Tree-based model for breast cancer prognostication

De-escalation of Axillary Surgery After Neoadjuvant Therapy

Surgical de-escalation of the axilla has evolved over the past 28 years since the emergence of sentinel lymph node surgery. Well-documented complications of the once standard of care axillary lymph node dissection (ALND), including lymphedema, led physician scientists towards a progressive push to study and incorporate less invasive techniques in the axilla. Many trials have justified oncologic safety of axillary de-escalation in patients who are spared neoadjuvant treatment. The applicability in the neoadjuvant setting, however, is less clear and axillary surgical approaches in this patient population have evolved at a slower pace. This review aims to analyze current data in axillary management for patients undergoing neoadjuvant treatment and to discuss current surgical approaches based on nodal pathologic response.

Application of Texture and Volume Model Analysis to Dedicated Axillary High-resolution 3D T2-weighted MR Imaging: A Novel Method for Diagnosing Lymph Node Metastasis in Patients with Clinically Node-negative Breast Cancer

PURPOSE

To evaluate the effectiveness of the texture analysis of axillary high-resolution 3D T2-weighted imaging (T2WI) in distinguishing positive and negative lymph node (LN) metastasis in patients with clinically node-negative breast cancer.

METHODS

Between December 2017 and May 2021, 242 consecutive patients underwent high-resolution 3D T2WI and were classified into the training (n = 160) and validation cohorts (n = 82). We performed manual 3D segmentation of all visible LNs in axillary level I to extract the texture features. As the additional parameters, the number of the LNs and the total volume of all LNs for each case were calculated. The least absolute shrinkage and selection operator algorithm and Random Forest were used to construct the models. We constructed the texture model using the features from the LN with the largest least axis length in the training cohort. Furthermore, we constructed the 3 models combining the selected texture features of the LN with the largest least axis length, the number of LNs, and the total volume of all LNs: texture-number model, texture-volume model, and texture-number-volume model. As a conventional method, we manually measured the largest cortical diameter. Moreover, we performed the receiver operating curve analysis in the validation cohort and compared area under the curves (AUCs) of the models.

RESULTS

The AUCs of the texture model, texture-number model, texture-volume model, texture-number-volume model, and conventional method in the validation cohort were 0.7677, 0.7403, 0.8129, 0.7448, and 0.6851, respectively. The AUC of the texture-volume model was higher than those of other models and conventional method. The sensitivity, specificity, positive predictive value, and negative predictive value of the texture-volume model were 90%, 69%, 49%, and 96%, respectively.

CONCLUSION

The texture-volume model of high-resolution 3D T2WI effectively distinguished positive and negative LN metastasis for patients with clinically node-negative breast cancer.

Clinical application of 18F-fluorodeoxyglucose positron emission tomography/computed tomography radiomics-based machine learning analyses in the field of oncology

Machine learning (ML) analyses using 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography (PET)/computed tomography (CT) radiomics features have been applied in the field of oncology. The current review aimed to summarize the current clinical articles about 18F-FDG PET/CT radiomics-based ML analyses to solve issues in classifying or constructing prediction models for several types of tumors. In these studies, lung and mediastinal tumors were the most commonly evaluated lesions, followed by lymphatic, abdominal, head and neck, breast, gynecological, and other types of tumors. Previous studies have commonly shown that 18F-FDG PET radiomics-based ML analysis has good performance in differentiating benign from malignant tumors, predicting tumor characteristics and stage, therapeutic response, and prognosis by examining significant differences in the area under the receiver operating characteristic curves, accuracies, or concordance indices (> 0.70). However, these studies have reported several ML algorithms. Moreover, different ML models have been applied for the same purpose. Thus, various procedures were used in 18F-FDG PET/CT radiomics-based ML analysis in oncology, and 18F-FDG PET/CT radiomics-based ML models, which are easy and universally applied in clinical practice, would be expected to be established.

Future Directions in the Assessment of Axillary Lymph Nodes in Patients with Breast Cancer

Background and Objectives: Breast cancer (BC) is a leading cause of morbidity and mortality worldwide, and accurate assessment of axillary lymph nodes (ALNs) is crucial for patient management and outcomes. We aim to summarize the current state of ALN assessment techniques in BC and provide insights into future directions. Materials and Methods: This review discusses various imaging techniques used for ALN evaluation, including ultrasound, computed tomography, magnetic resonance imaging, and positron emission tomography. It highlights advancements in these techniques and their potential to improve diagnostic accuracy. The review also examines landmark clinical trials that have influenced axillary management, such as the Z0011 trial and the IBCSG 23-01 trial. The role of artificial intelligence (AI), specifically deep learning algorithms, in improving ALN assessment is examined. Results: The review outlines the key findings of these trials, which demonstrated the feasibility of avoiding axillary lymph node dissection (ALND) in certain patient populations with low sentinel lymph node (SLN) burden. It also discusses ongoing trials, including the SOUND trial, which investigates the use of axillary ultrasound to identify patients who can safely avoid sentinel lymph node biopsy (SLNB). Furthermore, the potential of emerging techniques and the integration of AI in enhancing ALN assessment accuracy are presented. Conclusions: The review concludes that advancements in ALN assessment techniques have the potential to improve patient outcomes by reducing surgical complications while maintaining accurate disease staging. However, challenges such as standardization of imaging protocols and interpretation criteria need to be addressed. Future research should focus on large-scale clinical trials to validate emerging techniques and establish their efficacy and cost-effectiveness. Over-all, this review provides valuable insights into the current status and future directions of ALN assessment in BC, highlighting opportunities for improving patient care.

Predicting the risk of axillary lymph node metastasis in early breast cancer patients based on ultrasonographic-clinicopathologic features and the use of nomograms: a prospective single-center observational study

OBJECTIVES

The purpose of this study was to establish two preoperative nomograms to evaluate the risk for axillary lymph node (ALN) metastasis in early breast cancer patients based on ultrasonographic-clinicopathologic features.

METHODS

We prospectively evaluated 593 consecutive female participants who were diagnosed with cT_1-3N_0-1M₀ breast cancer between March 2018 and May 2019 at Sun Yat-Sen Memorial Hospital. The participants were randomly classified into training and validation sets in a 4:1 ratio for the development and validation of the nomograms, respectively. Multivariate logistic regression analysis was performed to identify independent predictors of ALN status. We developed Nomogram A and Nomogram B to predict ALN metastasis (presence vs. absence) and the number of metastatic ALNs (≤ 2 vs. > 2), respectively.

RESULTS

A total of 528 participants were evaluated in the final analyses. Multivariable analysis revealed that the number of suspicious lymph nodes, long axis, short-to-long axis ratio, cortical thickness, tumor location, and histological grade were independent predictors of ALN status. The AUCs of nomogram A in the training and validation groups were 0.83 and 0.78, respectively. The AUCs of nomogram B in the training and validation groups were 0.87 and 0.87, respectively. Both nomograms were well-calibrated.

CONCLUSION

We developed two preoperative nomograms that can be used to predict ALN metastasis (presence vs. absence) and the number of metastatic ALNs (≤ 2 vs. > 2) in early breast cancer patients. Both nomograms are useful tools that will help clinicians predict the risk of ALN metastasis and facilitate therapy decision-making about axillary surgery.

KEY POINTS

• We developed two preoperative nomograms to predict axillary lymph node status based on ultrasonographic-clinicopathologic features. • Nomogram A was used to predict axillary lymph node metastasis (presence vs. absence). The AUCs in the training and validation groups were 0.83 and 0.78, respectively. Nomogram B was used to estimate the number of metastatic lymph nodes ( ≤ 2 vs. > 2). The AUCs in the training and validation group were 0.87 and 0.87, respectively. • Our nomograms may help clinicians weigh the risks and benefits of axillary surgery more appropriately.

Prognostic factors for cancer-specific survival in 220 patients with breast cancer: A single center experience

OBJECTIVES

Hospital-based breast cancer survival studies are scarce in western Iran. Furthermore, the relationship between breast cancer survival and clinical parameters has been extensively studied, but many of the findings come from developing countries. This paper aims to estimate the survival of hospital-based breast cancer patients and its predictor factors.

METHOD

This retrospective analysis was conducted on 578 patients with primary breast cancer who underwent surgery between 2004 and 2020. Information was collected from medical reports by the Hospital information system in Imam Reza Hospital, Kermanshah, Iran. One-, 2-, 5-, and 10-year breast cancer-specific survival has been calculated using the Kaplan-Meier process. Crude and adjusted Hazard Ratios (HR) were calculated using the Cox proportional regression model.

RESULT

One-, 2-, and 5-year overall breast cancer survival were 219 (99.54%), 196 (89.09%), 159 (72.27%), and 70 (31.81%), respectively. Univariate analysis of breast cancer patients with tumor-related variables revealed that factors such as age, menopause status, lymph node metastasis, number of lymph nodes, organ metastasis, and stage of disease were significantly associated with disease-specific survival (p < .05). Multivariate analysis demonstrated that metastasis (HR = 41.77, 95% CI: 15.3-114.15) and lymph node metastasis (HR = 5.26, 95% CI: 1.9-14.6) were significantly related to survival.

CONCLUSION

The findings demonstrate that survival is relatively low and is consistent with late-stage disease diagnosis. It is believed that this is due to a poor level of awareness, lack of screening programs, and subsequent late access to treatment.

Development of High-Resolution Dedicated PET-Based Radiomics Machine Learning Model to Predict Axillary Lymph Node Status in Early-Stage Breast Cancer

Simple Summary

Accurate clinical axillary evaluation plays an important role in the diagnosis of and treatment planning for breast cancer (BC). This study aimed to develop a machine learning model integrating dedicated breast PET and clinical characteristics for prediction of axillary lymph node status in cT1-2N0-1M0 BC non-invasively. The performance of this integrating model in identifying pN0 and pN1 with the AUC was 0.94. We achieved an NPV of 96.88% in the cN0 and PPV of 92.73% in the cN1 subgroup. The higher true positive and true negative rate could delineate clinical subtypes and apply more precise treatment for patients with early-stage BC.

Abstract

Purpose of the Report: Accurate clinical axillary evaluation plays an important role in the diagnosis and treatment planning for early-stage breast cancer (BC). This study aimed to develop a scalable, non-invasive and robust machine learning model for predicting of the pathological node status using dedicated-PET integrating the clinical characteristics in early-stage BC. Materials and Methods: A total of 420 BC patients confirmed by postoperative pathology were retrospectively analyzed. 18F-fluorodeoxyglucose (¹⁸F-FDG) Mammi-PET, ultrasound, physical examination, Lymph-PET, and clinical characteristics were analyzed. The least absolute shrinkage and selection operator (LASSO) regression analysis were used in developing prediction models. The characteristic curve (ROC) of the area under receiver-operator (AUC) and DeLong test were used to evaluate and compare the performance of the models. The clinical utility of the models was determined via decision curve analysis (DCA). Then, a nomogram was developed based on the model with the best predictive efficiency and clinical utility and was validated using the calibration plots. Results: A total of 290 patients were enrolled in this study. The AUC of the integrated model diagnosed performance was 0.94 (95% confidence interval (CI), 0.91–0.97) in the training set (n = 203) and 0.93 (95% CI, 0.88–0.99) in the validation set (n = 87) (both p < 0.05). In clinical N0 subgroup, the negative predictive value reached 96.88%, and in clinical N1 subgroup, the positive predictive value reached 92.73%. Conclusions: The use of a machine learning integrated model can greatly improve the true positive and true negative rate of identifying clinical axillary lymph node status in early-stage BC.

Evaluation of the Histopathological Features of Early-stage Invasive Ductal Breast Carcinoma by 18Fluoride-fluorodeoxyglucose Positron Emission Tomography/Computed Tomography

Objectives:

This study investigates the relationship between ¹⁸fluoride-fluorodeoxyglucose (¹⁸F-FDG) positron emission tomography/computed tomography (PET/CT) parameters and histopathological features in patients with early-stage invasive ductal breast carcinoma (IDBC).

Methods:

Patients with early-stage IDBC who underwent ¹⁸F-FDG PET/CT scan for staging were included in this retrospective study. The status of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor-2, Ki-67 proliferative index, and grades of tumors were recorded. The conventional metabolic parameters [maximum standard uptake value (SUV_max) and average standard uptake value] and volume-based parameters [metabolic tumor volume (MTV) and total lesion glycolysis] of the primary tumor were obtained from the ¹⁸F-FDG PET/CT images. The associations and correlations between the ¹⁸F-FDG PET/CT parameters and histopathological features were assessed.

Results:

One hundred forty-three patients were included. ¹⁸F-FDG PET/CT parameters, other than MTV, were significantly associated with the ER and PR status and Ki-67 index, while T-staging was significantly associated with all ¹⁸F-FDG PET/CT parameters. In the axillary lymph node (ALN) involvement, no significant difference was found in the ¹⁸F-FDG PET/CT parameters. In terms of the pathological stage, a significant difference was found in all ¹⁸F-FDG PET/CT parameters. ¹⁸F-FDG PET/CT parameters, other than MTV, were significantly higher in non-luminal breast tumors than luminal tumors and in high-grade tumors than low-grade ones. Triple-negative tumors had the highest ¹⁸F-FDG PET/CT parameter, but the difference was insignificant for MTV. The SUV_max had the strongest correlation with Ki-67 proliferative index.

Conclusion:

Tumors with aggressive histopathological features had higher ¹⁸F-FDG PET/CT parameter values. This study suggests that ¹⁸F-FDG PET/CT may provide prognostic information in patients with early-stage IDBC.

Is ultrasound-guided fine needle aspiration biopsy of axillary lymph nodes a viable alternative to sentinel lymph node biopsy?

BACKGROUND

Axillary lymph node (ALN) ultrasound-guided fine needle aspiration biopsy (US-FNAB), a minimally invasive procedure, may be used for the preoperative evaluation of ALN status of breast cancer patients. Despite the relative ease of use and low cost, paucity of comparative studies and variation in the reported sensitivity of FNAB preclude its clinical utility in evaluation of ALNs. This study aims to determine the accuracy of US-FNAB in detecting metastasis in ALN pre-operatively and to assess US-FNAB as a viable alternative to sentinel lymph node (SLN) excision.

METHODS

The 228 consecutive ALN US-FNABs with subsequent histologic follow up performed from 2005 to 2020 in patients with breast carcinoma were retrospectively evaluated. FNAB results were correlated with histologic diagnosis. Sensitivity, specificity, accuracy, and risk of malignancy of FNAB were calculated.

RESULTS

157/228 (69%) FNABs were concordant with histology, 37/228 (16%) discordant. Positive FNAB findings correlated with primary tumor size, grade, number of metastatic lymph nodes and size of metastases. FNAB with negative diagnosis carried a 22% risk of malignancy, atypical 43%, suspicious 80%, and positive a 100% risk of malignancy (100% positive predictive value [PPV]). The sensitivity and specificity were 78% and 95% respectively; accuracy was 77%. SLN biopsy was avoided in all 82 (36%) cases with positive FNAB results.

CONCLUSION

Negative FNAB result does not exclude metastatic carcinoma. With 100% PPV, full ALN dissection and/or neoadjuvant chemotherapy can be safely planned after a positive FNAB result, avoiding SLN biopsy, reducing management costs and shortening time interval to definitive therapy.

A machine learning-based radiomics model for the prediction of axillary lymph-node metastasis in breast cancer

OBJECTIVE

The aim of this study was to develop and validate machine learning-based radiomics model for predicting axillary lymph-node (ALN) metastasis in invasive ductal breast cancer (IDC) using F-18 fluorodeoxyglucose (FDG) positron emission tomography/computed tomography (PET/CT).

METHODS

A total of 100 consecutive IDC patients who underwent surgical resection of primary tumor with sentinel lymph-node biopsy and/or ALN dissection without any neoadjuvant treatment were analyzed. Volume of interests (VOIs) were drawn more than 2.5 of standardized uptake value in the primary tumor on the PET scan using 3D slicer. Pyradiomics package was used for the extraction of texture features in python. The radiomics prediction model for ALN metastasis was developed in 75 patients of the training cohort and validated in 25 patients of the test cohort. XGBoost algorithm was utilized to select features and build radiomics model. The sensitivity, specificity, and accuracy of the predictive model were calculated.

RESULTS

ALN metastasis was found in 43 patients (43%). The sensitivity, specificity, and accuracy of F-18 FDG PET/CT for the diagnosis of ALN metastasis in the entire patients were 55.8%, 93%, and 77%, respectively. The radiomics model for the prediction of ALN metastasis was successfully developed. The sensitivity, specificity, and accuracy of the radiomics model for the prediction of ALN metastasis in the test cohorts were 90.9%, 71.4%, and 80%, respectively.

CONCLUSION

The machine learning-based radiomics model showed good sensitivity for the prediction of ALN metastasis and could assist the preoperative individualized prediction of ALN status in patients with IDC.

To Evaluate the Accuracy of Axillary Staging Using Ultrasound and Ultrasound-Guided Fine-Needle Aspiration Cytology (USG-FNAC) in Early Breast Cancer Patients-a Prospective Study

In breast cancer, axillary lymph node involvement directly impacts the patient survival and prognosis. Sentinel lymph node biopsy (SLNB) is a procedure of choice for axillary staging in early breast cancer. Currently, management options for axilla management are axillary lymph node dissection and sentinel node biopsy in node positive and in node negative respectively. Accuracy of current clinical methods for evaluating axilla is low. Hence, to select patients for appropriate procedure, ultrasound (USG) combined with fine-needle aspiration cytology (USG-FNAC) using vascular pedicle-based nodal mapping method is emerging as a good tool to address above issues. We evaluated the feasibility of ultrasound and needle aspiration cytology in a tertiary care center. All early breast cancer patients with clinically node-negative axilla and having palpable nodes with less than or equal to 5 cm tumor size in breast were screened by ultrasound of axilla to categorize the nodes as suspicious or non-suspicious based on radiological features and vascular pedicle-based nodal mapping method of axilla. Patients having suspicious nodes underwent ultrasound of axilla and needle aspiration; if found positive, patient underwent axillary node dissection. Sentinel node biopsy (SLNB) performed in all patients found negative on needle aspiration and in all patients having non-suspicious nodes on ultrasound axilla. Final histopathology was taken as gold standard. The sensitivity, specificity, accuracy, positive predictive value, and negative predictive value were calculated for ultrasound (USG) and ultrasound-guided needle aspiration (USG-FNAC). A total of 100 patients were included in which 58 had non-suspicious and 42 had suspicious nodes on ultrasound of axilla. Among suspicious group, 24 were positive on ultrasound-guided needle aspiration cytology and 18 were negative. In non-suspicious nodes, sentinel node biopsy was performed. Sensitivity, specificity, positive predictive value, and negative predictive value for ultrasound were 61.5%, 75.6%, 69.5%, and 68.5% respectively. For ultrasound-guided needle aspiration (USG-FNAC), sensitivity, specificity, and positive and negative predictive value are 83%, 100%, 100%, and 72.6% respectively. The accuracy of ultrasound (USG) and ultrasound-guided needle aspiration (USG-FNAC) was 69% and 88.1%. The result of our study indicates the feasibility of USG and USG-FNAC in a high-volume center with good accuracy of around 70-80%. Approximately one-fourth (24%) of the total patients were taken up for axillary lymph node dissection (ALND) without performing SLNB.

Maximum slope of ultrafast dynamic contrast-enhanced MRI of the breast: Comparisons with prognostic factors of breast cancer

PURPOSE

To determine the relationship between the maximum slope (MS) of ultrafast dynamic contrast-enhanced (DCE)-MRI and prognostic factors of breast cancer.

METHODS

One hundred thirteen patients with 118 breast cancers were included in this study. The ultrafast DCE sequence was acquired using a higher parallel imaging factor. Its spatial resolution was 0.9 × 0.9 × 2.5 mm and its temporal resolution was 8.3 s/phase. Each lesion was automatically segmented, and the ROI of highest enhancement in the lesion was identified. In this ROI, the MS was calculated. The MS of each lesion was compared with various prognostic factors of breast cancer.

RESULTS

The MS of invasive cancer (median: 9.81%/sec) was significantly higher than that of ductal carcinoma in situ (median: 7.26%/sec) (p = 0.001). In the ROC analysis, the area under the ROC curve (AUC) was 0.7295. The MS of invasive cancer with axillary lymph node (LN) metastasis (median: 11.97%/sec) was significantly higher than that without axillary LN metastasis (median: 9.425%/sec) (p = 0.0024). In the ROC analysis, the AUC was 0.7177. In addition, the MS became significantly higher as the level of the proliferation marker ki-67 increased (correlation coefficient: 0.3317) (p = 0.0009).

CONCLUSIONS

MS of ultrafast DCE-MRI is useful for predicting the prognostic factors of breast cancer. Higher maximum slope (MS) is significantly associated with an invasive breast cancer component. Higher MS is significantly associated with an axillary lymph node metastasis. MS becomes significantly higher with increasing ki-67 (a proliferation marker). Ultrafast MRI is useful for predicting the prognostic factors of breast cancer.

Assessment of Metastatic and Reactive Sentinel Lymph Nodes with B7-H3-Targeted Ultrasound Molecular Imaging: A Longitudinal Study in Mouse Models

PURPOSE

To explore the potential of B7-H3-targeted ultrasound molecular imaging (USMI) for longitudinal assessment and differentiation of metastatic and reactive sentinel lymph nodes (SLNs) in mouse models.

PROCEDURES

Metastatic and reactive SLN models were established by injection of 4T1 breast cancer cells and complete Freund's adjuvant (CFA) respectively to the 4th mammary fat pad of female BALB/c mice. At day 21, 28, and 35 after inoculation, USMI was performed following intravenous injection of B7-H3-targeted microbubbles (MBB7-H3) or IgG-control microbubbles (MBcontrol). All SLNs were histopathologically examined after the last imaging session.

RESULTS

A total of 20 SLNs from tumor-bearing mice (T-SLNs) and five SLNs from CFA-injected mice (C-SLNs) were examined by USMI. Nine T-SLNs were histopathologically positive for metastasis (MT-SLNs). From day 21 to 35, T-SLNs showed a rising trend in MBB7-H3 signal with a steep increase in MT-SLNs at day 35 (213.5 ± 80.8 a.u.) as compared to day 28 (87.6 ± 77.2 a.u., P = 0.002) and day 21 (55.7 ± 35.5 a.u., P < 0.001). At day 35, MT-SLNs had significantly higher MBB7-H3 signal than non-metastatic T-SLNs (NMT-SLNs) (101.9 ± 48.0 a.u., P = 0.001) and C-SLNs (38.5 ± 34.0 a.u., P = 0.001); MBB7-H3 signal was significantly higher than MBcontrol in MT-SLNs (P = 0.001), but not in NMT-SLNs or C-SLNs (both P > 0.05). A significant correlation was detected between MBB7-H3 signal and volume fraction of metastasis in MT-SLNs (r = 0.76, P = 0.017).

CONCLUSIONS

B7-H3-targeted USMI allows differentiation of MT-SLNs from NMT-SLNs and C-SLNs in mouse models and has great potential to evaluate tumor burden in SLNs of breast cancer.

Estrogen Receptor Positive and Progesterone Receptor Negative Breast Cancer: the Role of Hormone Therapy

ER+/PR- (estrogen receptor positive and progesterone receptor negative) tumors constitute only a small portion of the breast cancer population. Patients with ER+/PR- tumors, however, are characterized by worse survival compared to patients with ER+/PR+ (estrogen receptor positive and progesterone receptor positive) tumors. Controversy exists regarding the efficacy of hormone blocking therapy for patients with ER+/PR- tumors. The NCDB was queried between 2004 and 2015, and patients with invasive ER+/PR- tumors were identified. We employed univariate Cox proportional hazards to compare outcomes among patients that did or did not receive hormone blocking therapy. We identified 138,398 patients with invasive ER+/PR- tumors, 32,044 (23%) of whom did not receive hormone blocking therapy. The reasons for not receiving hormone blocking therapy included contraindications to treatment, death, patient refusal, and unknown. There were no significant differences in race, income quartile, or education quartile between patients who did and did not receive hormone blocking therapy. Patients who did not receive hormone blocking therapy underwent surgical assessment of the axilla more frequently than those who did receive hormone therapy. Our analysis demonstrated that hormone blocking therapy administration was associated with increased overall survival for up to 10 years of follow up (HR: 0.58; 95% CI: 0.56-0.59, p < 0.001). Hormone blocking therapy may be associated with increased survival for breast cancer patients with ER+/PR- tumors. Although this benefit may last for years after completion of the course, up to 25% of patients do not receive this treatment. Strategies to increase the utilization and adherence to hormone blocking therapy regimens may improve patient survival outcomes.

Tree-Based Analysis

Tree-based methods have become one of the most flexible, intuitive, and powerful data analytic tools for exploring complex data structures. Tree-based methods provide a natural framework for creating patient subgroups for risk classification. In this article, we review methodological and practical aspects of tree-based methods, with a focus on diagnostic classification (binary outcome) and prognostication (censored survival outcome). Creating an ensemble of trees improves prediction accuracy and addresses instability in a single tree. Ensemble methods are described that rely on resampling from the original data. Finally, we present methods to identify a representative tree from the ensemble that can be used for clinical decision-making. The methods are illustrated using data on ischemic heart disease classification, and data from the SPRINT trial (Systolic Blood Pressure Intervention Trial) on adverse events in patients with high blood pressure.

Diagnostic performance of standard breast MRI compared to dedicated axillary MRI for assessment of node-negative and node-positive breast cancer

Objectives

To investigate whether breast MRI has comparable diagnostic performance as dedicated axillary MRI regarding assessment of node-negative and node-positive breast cancer.

Methods

Forty-seven patients were included. All had undergone both breast MRI and dedicated axillary MRI, followed by surgery. All included breast MRI exams had complete field of view (FOV) of the axillary region. First, unenhanced T2-weighted (T2W) and subsequent diffusion-weighted (DW) images of both MRI exams were independently analyzed by two breast radiologists using a confidence scale and compared to histopathology. ADC values were measured by two researchers independently. Diagnostic performance parameters were calculated on a patient-by-patient basis.

Results

T2W breast MRI had the following diagnostic performance: sensitivity of 50.0% and 62.5%, specificity of 92.3%, PPV of 57.1% and 62.5%, NPV of 90.0% and 92.3%, and AUC of 0.72 for reader 1 and 0.78 for reader 2. T2W dedicated axillary MRI had the following diagnostic performance: sensitivity of 37.5% and 62.5%, specificity of 82.1% and 92.3%, PPV of 44.6% and 50.0%, NPV of 87.8% and 91.4%, and AUC of 0.65 for reader 1 and 0.73 for reader 2. In both evaluations, addition of DW images resulted in comparable diagnostic performance. For both breast MRI and dedicated axillary MRI, there was no significant difference between mean ADC values of benign and malignant lymph nodes.

Conclusions

T2W breast MRI with complete FOV of the axillary region has comparable diagnostic performance as T2W dedicated axillary MRI regarding assessment of node-negative and node-positive breast cancer. Optimization of T2W breast MRI protocol by including a complete FOV of the axillary region can, therefore, be recommended in clinical practice.

Key Points

• Breast MRI with complete field of view of the axillary region has comparable diagnostic performance as dedicated axillary MRI regarding assessment of node-negative and node-positive breast cancer.

• Optimization of breast MRI protocol by including a complete field of view of the axillary region is recommended in clinical practice.

• For both breast MRI and dedicated axillary MRI, DW imaging (including ADC measurements) is of no added value.

Visceral fat metabolic activity evaluated by preoperative 18F-FDG PET/CT significantly affects axillary lymph node metastasis in postmenopausal luminal breast cancer

Obesity is known to increase breast cancer risk and aggressiveness in postmenopausal luminal breast cancer and obesity-driven dysfunctional metabolic activity in visceral adipose tissue (VAT) is considered as one of the principal underlying mechanism. We aimed to investigate the relationship between VAT metabolic activity evaluated by preoperative ¹⁸F-FDG PET/CT and axillary lymph node (ALN) metastasis in postmenopausal luminal breast cancer patients. In total, 173 patients were enrolled in study. They all underwent preoperative ¹⁸F-FDG PET/CT and surgery. VAT metabolic activity was defined as the maximum standardized uptake value (SUVmax) of VAT divided by the SUVmax of subcutaneous adipose tissue (V/S ratio). In luminal breast cancer, the patients with ALN metastasis showed significantly higher V/S ratio than the patients without ALN metastasis. Furthermore, V/S ratio was significantly associated with ALN metastasis in luminal breast cancer patients. Erythrocyte sedimentation rate, which reflect the systemic inflammation, was significantly higher in ALN metastasis group than the negative ALN metastasis group in luminal breast cancer patients and showed significant positive correlation with V/S ratio. V/S ratio significantly affects the ALN metastasis status in postmenopausal luminal breast cancer patients and it may be useful as a potential biomarker of obesity-driven systemic inflammation associated with tumor aggressiveness.

Diagnostic performance of axillary ultrasound and standard breast MRI for differentiation between limited and advanced axillary nodal disease in clinically node-positive breast cancer patients

Preoperative differentiation between limited (pN1; 1-3 axillary metastases) and advanced (pN2-3; ≥4 axillary metastases) nodal disease can provide relevant information regarding surgical planning and guiding adjuvant radiation therapy. The aim was to evaluate the diagnostic performance of preoperative axillary ultrasound (US) and breast MRI for differentiation between pN1 and pN2-3 in clinically node-positive breast cancer. A total of 49 patients were included with axillary metastasis confirmed by US-guided tissue sampling. All had undergone breast MRI between 2008-2014 and subsequent axillary lymph node dissection. Unenhanced T2-weighted MRI exams were reviewed by two radiologists independently. Each lymph node on the MRI exams was scored using a confidence scale (0-4) and compared with histopathology. Diagnostic performance parameters were calculated for differentiation between pN1 and pN2-3. Interobserver agreement was determined using Cohen's kappa coefficient. At final histopathology, 67.3% (33/49) and 32.7% (16/49) of patients were pN1 and pN2-3, respectively. Breast MRI was comparable to US in terms of accuracy (MRI reader 1 vs US, 71.4% vs 69.4%, p = 0.99; MRI reader 2 vs US, 73.5% vs 69.4%, p = 0.77). In the case of 1-3 suspicious lymph nodes, pN2-3 was observed in 30.4% on US (positive predictive value (PPV) 69.6%) and in 22.2-24.3% on MRI (PPV 75.7-77.8%). In the case of ≥4 suspicious lymph nodes, pN1 was observed in 33.3% on US (negative predictive value (NPV) 66.7%) and in 38.5-41.7% on MRI (NPV 58.3-61.5%). Interobserver agreement was considered good (k = 0.73). In clinically node-positive patients, the diagnostic performance of axillary US and breast MRI is comparable and limited for accurate differentiation between pN1 and pN2-3. Therefore, there seems no added clinical value of preoperative breast MRI regarding nodal staging in patients with positive axillary US.

Developing and Validating Risk Assessment Models of Clinical Outcomes in Modern Oncology

The identification of prognostic factors and building of risk assessment prognostic models will continue to play a major role in 21st century medicine in patient management and decision making. Investigators are often interested in examining the relationship between host, tumor-related, and environmental variables in predicting clinical outcomes. We make a distinction between static and dynamic prediction models. In static prediction modelling, typically variables collected at baseline are utilized in building models. On the other hand, dynamic predictive models leverage the longitudinal data of covariates collected during treatment or follow-up, and hence provide accurate predictions of patients prognoses. To date, most risk assessment models in oncology have been based on static models. In this article, we cover topics that are related to the analysis of prognostic factors, centering on factors that are both relevant at the time of diagnosis or initial treatment and during treatment. We describe the types of risk prediction and then provide a brief description of the penalized regression methods. We then review the state-of-the art methods for dynamic prediction and compare the strengths and the limitations of these methods. While static models will continue to play an important role in oncology, developing and validating dynamic models of clinical outcomes need to take a higher priority. It is apparent that a framework for developing and validating dynamic tools in oncology is still needed. One of the limitations in oncology that modelers may be constrained by the lack of access to the longitudinal biomarker data. It is highly recommended that the next generation of risk assessments consider the longitudinal biomarker data and outcomes so that prediction can be continually updated.

Long-term prognostic effect of hormone receptor subtype on breast cancer

PURPOSE

To determine the long-term prognostic role of hormone receptor subtype in breast cancer using surveillance, epidemiology, and end results (SEER) database.

METHODS

Data of 810,587 female operable invasive breast cancer patients from SEER database with a mean follow-up period of 94.2 months (range, 0-311 months) were analyzed. Hormone receptor subtype was classified into four groups based on estrogen receptor (ER) and progesterone receptor (PR) statuses: ER(+)/PR(+), ER(+)/PR(-), ER(-)/PR(+), and ER(-)/PR(-).

RESULTS

Numbers of subjects with ER(+)/PR(+), ER(+)/PR(-), ER(-)/PR(+), ER(-)/PR(-), and unknown were 496,279 (61.2%), 86,858 (10.7%), 11,545 (1.4%), 135,441 (16.7%), and 80,464 (9.9%), respectively. The ER(+)/PR(+) subtype showed the best breast-cancer-specific survival, followed by ER(+)/PR(-), ER(-)/PR(+), and ER(-)/PR(-) subtypes in the respective order (all p < 0.001). Survival difference among hormone receptor subtypes was maintained in subgroup analysis according to anatomic stage, race, age group, and year of diagnosis. Hormone receptor subtype was a significant independent prognostic factor in multivariable analyses (p < 0.001). Hazard ratios of ER(+)/PR(-), ER(-)/PR(+), and ER(-)/PR(-) for breast-cancer-specific mortality risk were 1.419 (95% confidence interval [CI] 1.383-1.456), 1.630 (95% CI 1.537-1.729), and 1.811 (95% CI 1.773-1.848), respectively, with ER(+)/PR(+) as reference.

CONCLUSION

Hormone receptor subtype is a significant independent prognostic factor in female operable invasive breast cancer patients with long-term effect. The ER(+)/PR(+) subtype shows the most favorable prognosis, followed by ER(+)/PR(-), ER(-)/PR(+), and ER(-)/PR(-) subtypes in the respective order. Prognostic impacts of hormone receptor subtypes are also maintained in subgroup analysis according to anatomic stage, race, age, and year of diagnosis.

Diagnostic value of magnetic resonance imaging using superparamagnetic iron oxide for axillary node metastasis in patients with breast cancer: a meta-analysis

Objectives Identification of axillary metastases in breast cancer is important for staging disease and planning treatment, but current techniques are associated with a number of adverse events. This report evaluates the diagnostic accuracy of superparamagnetic iron oxide (SPIO)-enhanced magnetic resonance imaging (MRI) techniques for identification of axillary metastases in breast cancer patients. Methods We performed a meta-analysis of previous studies that compared SPIO enhanced MRI with histological diagnosis after surgery or biopsy. We searched PubMed, Ovid, Springer Link, and Cochrane library to identify studies reporting data for SPIO enhanced MRI for detection of axillary lymph node metastases in breast cancer until December 2013. The following keywords were used: “magnetic resonance imaging AND axilla” and “superparamagnetic iron oxide AND axilla”. Eligible studies were those that compared SPIO enhanced MRI with histological diagnosis. Sensitivity and specificity were calculated for every study; summary receiver operating characteristic and subgroup analyses were done. Study quality and heterogeneity were also assessed. Results There were 7 publications that met the criteria for inclusion in our meta-analysis. SROC curve analysis for per patient data showed an overall sensitivity of 0.83 (95% Confidence interval (CI): 0.75–0.89) and overall specificity of 0.97 (95% CI: 0.94–0.98). Overall weighted area under the curve was 0.9563. Conclusions SPIO enhanced MRI showed a trend toward high diagnostic accuracy in detection of lymph node metastases for breast cancer. So, when the breast cancer patients has axillary metastases histologically, SPIO enhanced MRI may be effective diagnostic imaging modality for axillary metastases.

Perfusion contrast-enhanced ultrasound to predict early lymph-node metastasis in breast cancer

PURPOSE

To evaluate whether quantitative analysis of perfusion contrast-enhanced ultrasound (CE-US) could predict early lymph-node (LN) metastasis in clinically node-negative breast cancer.

MATERIALS AND METHODS

In this prospective study, 64 breast cancer patients were selected for perfusion CE-US imaging. Regions of interest were placed where the strongest and weakest signal increases were found to obtain peak intensities (PIs; PI_max and PI_min, respectively) for time-intensity curve analyzes. The PI difference and PI ratio were calculated as follows: PI difference = PI_max-PI_min; PI ratio = PI_max/PI_min.

RESULTS

Forty-seven cases were histologically diagnosed as negative for LN metastasis and 17 were positive. There was a significant difference in PI_min and the PI ratio between the LN-negative and -positive metastasis groups (p = 0.0053 and 0.0082, respectively). Receiver-operating curve analysis revealed that the area under the curve of PI_min and the PI ratio were 0.73 and 0.72, respectively. The most effective threshold for the PI ratio was 1.52, and the sensitivity, specificity, positive predictive value, and negative predictive value were 59% (10/17), 87% (41/47), 63% (10/16), and 85% (41/48), respectively.

CONCLUSIONS

Parameters from the quantitative analysis of perfusion CE-US imaging showed significant differences between the LN-negative and -positive metastasis groups in clinically node-negative breast cancer.

Treatment-Free Survival in Patients With Differentiated Thyroid Cancer

OBJECTIVE

Cancer recurrence is a primary concern for patients with differentiated thyroid cancer; however, population-level data on recurrent or persistent disease do not currently exist. The objective of this study was to determine treated recurrent or persistent thyroid cancer by using a population-based registry, identify correlates of poor treatment-free survival, and define prognostic groups for treatment-free survival.

METHODS

In this population-based study, we evaluated treatment-free survival in 9273 patients from the Surveillance, Epidemiology, and End Results Program-Medicare with a diagnosis of differentiated thyroid cancer between 1998 and 2012. Treated recurrence was defined by treatment of recurrent or persistent differentiated thyroid cancer with surgery, radioactive iodine, or radiation therapy at ≥1 year after diagnosis. Multivariable analysis was performed with Cox proportional hazards regression, survival trees, and random survival forests.

RESULTS

In this cohort the median patient age at time of diagnosis was 69 years, and 75% of the patients were female. Using survival tree analyses, we identified five distinct prognostic groups (P < 0.001), with a prediction accuracy of 88.7%. The 5-year treatment-free survival rates of these prognostic groups were 96%, 91%, 85%, 72%, and 52%, respectively, and the 10-year treatment-free survival rates were 94%, 87%, 80%, 64%, and 39%. Based on survival forest analysis, the most important factors for predicting treatment-free survival were stage, tumor size, and receipt of radioactive iodine.

CONCLUSION

In this population-based cohort, five prognostic groups for treatment-free survival were identified. Understanding treatment-free survival has implications for the care and long-term surveillance of patients with differentiated thyroid cancer.

Analytical and clinical performance of progesterone receptor antibodies in breast cancer

OBJECTIVE

Comparison of analytical and immunohistochemical performance of progesterone receptor (PR) antibodies with correlation to recurrence of invasive breast cancer treated with endocrine therapy.

METHODS

The binding-affinity kinetics of PR clones 1E2, 1A6, 16 and 636 were compared using synthetic peptides derived from identified epitopes on a Biacore T200. A cohort of 351 cases (Hormone Receptor (HR)+/HER2-) were stained for PR expression with immunohistochemistry (IHC) and scored according to ASCO/CAP criteria.

RESULTS

The stability of the antigen/antibody complex was greater for the 1E2 clone compared to 1A6, 16 and 636 clones. PR IHC on archival tissue resulted in 94.3% (299/317) concordance with clones.

CONCLUSION

Clones evaluated in this study had a high level of concordance with IHC despite PR (1E2) demonstrating higher analytical binding properties than other clones. In a minority of cases (1.3% for 1E2 and 2.5% for 636) IHC results could convert estrogen receptor (ER)-/PR- to ER-/PR+ tumors, making these patients potentially eligible for endocrine therapy.

MicroRNA regulation of progesterone receptor in breast cancer

Hormone receptor status is of significant value when deciding on anti-estrogenic adjuvant therapy for breast cancer tumors. However, while estrogen receptor (ER) regulation was intensively studied, the regulation of progesterone receptor (PR) levels has not been extensively investigated. MicroRNAs (miRNAs, miRs) are post-transcriptional negative regulators of gene expression involved in diverse cellular processes. The aim of this study was to identify miRNAs that regulate PR in breast cancer.We mapped potential miRNA binding sites for miR-181a, miR-23a and miR-26b on PR mRNA and demonstrated a direct regulation of PR by these three miRNAs by in-vitro Luciferase binding assays. Over-expression of each miRNA in MCF-7 cells resulted in a reduction in the expression levels of PR mRNA. Then, expression levels of these miRNAs were measured in Formalin-Fixed, Paraffin-Embedded (FFPE) samples of 29 ER-positive breast cancer tumors and adjacent normal breast tissues. A significant reciprocal correlation between PR mRNA and the miRNA levels were identified suggesting a role for miR-181a, miR-23a and miR-26b in PR regulation in breast cancer. Moreover, the average expression fold-changes of the three miRNAs between cancerous and normal tissues displayed an opposite trend when analyzing according to Immuno-histochemistry(IHC) status. Furthermore, miR-181a and miR-26b were found to be over-expressed in most tumor tissues supporting their role in ER-positive breast cancer development. We conclude that miR-181a, miR-23a and miR-26b act as negative regulators of PR expression in ER-positive breast cancer. The diagnostic and prognostic potential of these miRNAs in breast cancer should be further evaluated.

A Prediction Model for Breast Cancer Recurrence after Adjuvant Hormone Therapy

Hormone therapy with tamoxifen has long been the established adjuvant treatment for node-positive, estrogen–receptor-positive breast cancer in postmenopausal women. Since 30–40% of these patients fail to respond, reliable outcome prediction is necessary for successful treatment allocation. Using pathobiological variables (available in most clinical records: tumor size, nodal involvement, estrogen and progesterone receptor content) from 596 patients recruited at a comprehensive cancer center, we developed a prediction model which we validated in an independent cohort of 175 patients recruited at a general hospital. Calculated at 3 and 4 years of follow-up, the discrimination indices were 0.716 [confidence limits (CL) 0.641, 0.752] and 0.714 (CL 0.650, 0.750) for the training data, and 0.726 (CL 0.591, 0.769) and 0.677 (CL 0.580, 0.745) for the testing data. Waiting for more effective approaches from genomic and proteomic studies, a model based on consolidated pathobiological variables routinely assessed at relatively low costs may be considered as the reference for assessing the gain of new markers over traditional ones, thus substantially improving the conventional use of prognostic criteria.

Development and validation of a multivariate predictive model for rheumatoid arthritis mortality using a machine learning approach

We developed and independently validated a rheumatoid arthritis (RA) mortality prediction model using the machine learning method Random Survival Forests (RSF). Two independent cohorts from Madrid (Spain) were used: the Hospital Clínico San Carlos RA Cohort (HCSC-RAC; training; 1,461 patients), and the Hospital Universitario de La Princesa Early Arthritis Register Longitudinal study (PEARL; validation; 280 patients). Demographic and clinical-related variables collected during the first two years after disease diagnosis were used. 148 and 21 patients from HCSC-RAC and PEARL died during a median follow-up time of 4.3 and 5.0 years, respectively. Age at diagnosis, median erythrocyte sedimentation rate, and number of hospital admissions showed the higher predictive capacity. Prediction errors in the training and validation cohorts were 0.187 and 0.233, respectively. A survival tree identified five mortality risk groups using the predicted ensemble mortality. After 1 and 7 years of follow-up, time-dependent specificity and sensitivity in the validation cohort were 0.79–0.80 and 0.43–0.48, respectively, using the cut-off value dividing the two lower risk categories. Calibration curves showed overestimation of the mortality risk in the validation cohort. In conclusion, we were able to develop a clinical prediction model for RA mortality using RSF, providing evidence for further work on external validation.

Predictive Value of 18F-FDG PET/CT for Axillary Lymph Node Metastasis in Invasive Ductal Breast Cancer

BACKGROUND

This study assessed whether primary tumor maximum standardized uptake value (pSUV_max) measured by ¹⁸F-fluoro-2-deoxy-D-glucose (¹⁸F-FDG) positron emission tomography/computed tomography (PET/CT) could improve the prediction of axillary lymph node (ALN) metastasis in invasive ductal breast cancer (IDC).

METHODS

In this study, 128 IDC patients who underwent pretreatment ¹⁸F-FDG PET/CT and surgical resection of primary tumor with sentinel lymph node biopsy, ALN dissection, or both were analyzed. All the patients were classified as five molecular subtypes. The optimal cutoff values of pSUVmax for all the patients and each molecular subtype for the prediction of ALN metastasis were determined using receiver operating characteristic (ROC) analysis. Furthermore, the prognostic accuracy of ALN metastasis was assessed using c-statistics.

RESULTS

The findings showed ALN metastasis in 52 patients (40.6%). The ¹⁸F-FDG PET/CT procedure had a sensitivity of 48.1% and a specificity of 94.7% for ALN metastasis. In the ROC analysis of pSUVmax for ALN metastasis, the optimal cutoff value was 3.9 for all the patients, 2.8 for the luminal A subtype, 3.3 for the luminal B (human epidermal growth factor receptor 2 [HER2]-negative) subtype, 5.3 for the luminal B (HER2-positive) subtype, 12.7 for the HER2-positive subtype, and 11.5 for the triple-negative subtype. A predictive ALN metastasis model using nodal ¹⁸F-FDG uptake finding gave a c-statistic of 0.714, and a model combination of nodal ¹⁸F-FDG uptake finding with pSUV_max of all the patients gave a c-statistic of 0.736 (P = 0.3926). However, the combination of nodal the ¹⁸F-FDG uptake finding with the pSUV_max of each molecular subtype gave a c-statistic of 0.791 (P = 0.0047).

CONCLUSIONS

Combining the pSUV_max of each molecular subtype with the nodal ¹⁸F-FDG uptake finding can improve the prediction of ALN metastasis in IDC.

Evaluation of three commercial progesterone receptor assays in a single tamoxifen-treated breast cancer cohort

Estrogen receptor and progesterone receptor status are routinely assessed using immunohistochemistry assays to assist in patient prognosis and clinical management. Three commonly utilized autostainer vendors-Dako, Leica and Ventana-provide ready-to-use progesterone receptor assays; however, they have never been directly compared in a single breast cancer cohort. We looked at three immunohistochemical progesterone receptor assays, in addition to original ligand-binding assay results, in a single retrospective, tamoxifen-treated breast cancer cohort to investigate inter- and intra-observer agreement, concordance, prognostic ability and measures of test performance. All immunohistochemical assays utilized the manufacturer's specified protocols. Five-year disease-free survival was the endpoint of interest, and multivariate models were adjusted for lymph node status, tumor grade, tumor size and human epidermal growth factor 2 status. All assays showed substantial to almost perfect agreement between the three observers (Dako κ=0.69-0.90; Leica κ=0.70-0.89; and Ventana κ=0.78-0.94) and concordance (Dako/Leica κ=0.81; Dako/Ventana κ=0.78; and Leica/Ventana κ=0.82). Univariate survival analyses showed that only the ligand-binding assay, Dako and Ventana assays achieved statistical significance. No statistically significant results were seen in multivariate models, although a strong trend was seen with the Ventana progesterone receptor assay. All assays performed similarly with regards to measures of test performance with ligand-binding assay set as the reference, and all immunohistochemical assays outperformed the ligand-binding assay in regards to 5-year disease-free survival. Despite similar agreement and concordance with the progesterone receptor assays, clear differences were noted with regards to 5-year disease-free survival. Additional survival analyses suggest that clinical utility of estrogen receptor assays vary when investigated in combination with progesterone receptor.

A nomogram to predict the probability of axillary lymph node metastasis in early breast cancer patients with positive axillary ultrasound

Among patients with a preoperative positive axillary ultrasound, around 40% of them are pathologically proved to be free from axillary lymph node (ALN) metastasis. We aimed to develop and validate a model to predict the probability of ALN metastasis as a preoperative tool to support clinical decision-making. Clinicopathological features of 322 early breast cancer patients with positive axillary ultrasound findings were analyzed. Multivariate logistic regression analysis was performed to identify independent predictors of ALN metastasis. A model was created from the logistic regression analysis, comprising lymph node transverse diameter, cortex thickness, hilum status, clinical tumour size, histological grade and estrogen receptor, and it was subsequently validated in another 234 patients. Coefficient of determination (R(2)) and the area under the ROC curve (AUC) were calculated to be 0.9375 and 0.864, showing good calibration and discrimination of the model, respectively. The false-negative rates of the model were 0% and 5.3% for the predicted probability cut-off points of 7.1% and 13.8%, respectively. This means that omission of axillary surgery may be safe for patients with a predictive probability of less than 13.8%. After further validation in clinical practice, this model may support increasingly limited surgical approaches to the axilla in breast cancer.

Preoperative axillary lymph node evaluation in breast cancer patients by breast magnetic resonance imaging (MRI): Can breast MRI exclude advanced nodal disease?

OBJECTIVES

To evaluate the diagnostic performance of breast magnetic resonance imaging (MRI) in preoperative evaluation of axillary lymph node metastasis (ALNM) in breast cancer patients and to assess whether breast MRI can be used to exclude advanced nodal disease.

METHODS

A total of 425 patients were included in this study and breast MRI findings were retrospectively reviewed. The diagnostic performance of breast MRI for diagnosis of ALNM was evaluated in all patients, patients with neoadjuvant chemotherapy (NAC), and those without NAC (no-NAC). We evaluated whether negative MRI findings (cN0) can exclude advanced nodal disease (pN2-pN3) using the negative predictive value (NPV) in each group.

RESULTS

The sensitivity and NPV of breast MRI in evaluation of ALNM was 51.3 % (60/117) and 83.3 % (284/341), respectively. For cN0 cases on MRI, pN2-pN3 manifested in 1.8 % (6/341) of the overall patients, 0.4 % (1/257) of the no-NAC group, and 6 % (5/84) of the NAC group. The NPV of negative MRI findings for exclusion of pN2-pN3 was higher for the no-NAC group than for the NAC group (99.6 % vs. 94.0 %, p = 0.039).

CONCLUSIONS

Negative MRI findings (cN0) can exclude the presence of advanced nodal disease with an NPV of 99.6 % in the no-NAC group.

KEY POINTS

• Breast MRI can be used to exclude advanced nodal disease (pN2-3). • Negative MRI allows breast cancer patients to avoid unnecessary axillary surgery (98.2 %). • Negative MRI findings exclude 99.6 % of pN2-pN3 in the no-NAC group. • Negative MRI findings exclude 96.0 % of pN2-pN3 in the NAC group.

Do All Women With Abnormal Sonographic Axillary Lymph Nodes Need a Biopsy?

PURPOSE

With the increased use of breast ultrasound for different indications, sonographically abnormal axillary lymph nodes are not a rare finding. We examined clinical and imaging characteristics in correlation with pathological reports of the sonographic guided biopsies to assess the yield of needle biopsy of these nodes.

METHODS

Clinical, imaging and pathology data were collected for 171 consecutive patients who underwent sonographic guided needle biopsy of an abnormal lymph node between 2008 and 2013. Malignancy rates were examined for different clinical settings: palpable axillary mass, previous history of breast cancer, findings suggestive of a systemic disease, and those with a breast finding of low suspicion or an incidental abnormal axillary lymph node. Patients with newly diagnosed breast cancer were excluded.

RESULTS

Twelve patients (7%) were found to have a malignancy on their axillary lymph node biopsy. Malignancy rates increased with age, and varied with clinical presentation: Axillary mass (8, 26%); history of breast cancer (2, 11%); systemic disease (0%) and breast finding of low suspicion or incidental abnormal lymph node on screening (1, 1%). Low rates of malignancy were found when the cortex was <6 mm (1, 0.8%). The most important imaging finding associated with malignancy was lack of a preserved hilum, in which case almost a third (10, 29%) of the biopsies were malignant. Only 1 of 89 women with a breast finding of low suspicion or an incidental abnormal axillary lymph node was found to have malignancy. In this case the lymph node had no hilum.

CONCLUSIONS

In women without breast cancer, a highly suspicious breast mass or an axillary mass, more stringent criteria should be used when evaluating an abnormal axillary lymph node on sonography, as the malignancy rates are very low (1%).

Application of Random Forest Survival Models to Increase Generalizability of Decision Trees: A Case Study in Acute Myocardial Infarction

Background. Tree models provide easily interpretable prognostic tool, but instable results. Two approaches to enhance the generalizability of the results are pruning and random survival forest (RSF). The aim of this study is to assess the generalizability of saturated tree (ST), pruned tree (PT), and RSF. Methods. Data of 607 patients was randomly divided into training and test set applying 10-fold cross-validation. Using training sets, all three models were applied. Using Log-Rank test, ST was constructed by searching for optimal cutoffs. PT was selected plotting error rate versus minimum sample size in terminal nodes. In construction of RSF, 1000 bootstrap samples were drawn from the training set. C-index and integrated Brier score (IBS) statistic were used to compare models. Results. ST provides the most overoptimized statistics. Mean difference between C-index in training and test set was 0.237. Corresponding figure in PT and RSF was 0.054 and 0.007. In terms of IBS, the difference was 0.136 in ST, 0.021 in PT, and 0.0003 in RSF. Conclusion. Pruning of tree and assessment of its performance of a test set partially improve the generalizability of decision trees. RSF provides results that are highly generalizable.

Diagnostic Accuracy and Impact on Management of Ultrasonography-Guided Fine-Needle Aspiration to Detect Axillary Metastasis in Breast Cancer Patients: A Prospective Study

BACKGROUND

The axillary nodal status is essential to determine the stage of disease at diagnosis. Our aim was to prospectively assess the diagnostic accuracy of ultrasonography-guided fine-needle aspiration (US-FNA) for the detection of metastasis in axillary lymph nodes in patients with breast cancer (BC) and its impact on the therapeutic decision.

MATERIALS AND METHODS

Ultrasonography (US) was performed in 407 axillae of 396 patients who subsequently underwent surgery. US-FNA was conducted when lymph nodes were detected by US. Axillary dissection (AD) was performed when US-FNA was positive for metastasis. Patients with negative US-FNA and breast tumors of 30 mm in size were candidates for selective sentinel lymph node biopsy (SLNB). The anatomopathological results of AD or SLNB were used as reference tests.

RESULTS

Lymph nodes were detected by US in 207 (50.8%) axillae. Of these, US-FNA was performed on 180 (86.9%). 94 axillae (52.2%) were positive for carcinoma and 79 women received AD. US-FNA had 77.5% sensitivity, 100% specificity, 100% positive predictive value, 69.3% negative predictive value, and 85.1% diagnostic accuracy. US-FNA avoided SLNB in 18.1% of patients who underwent AD.

CONCLUSIONS

Axillary US-FNA is an accurate technique in the staging of patients with BC. It allows reducing the number of SLNB and, when positive, offers a fast and useful tool.

Prognostic Significance of Single Progesterone Receptor Positivity: A Comparison Study of Estrogen Receptor Negative/Progesterone Receptor Positive/Her2 Negative Primary Breast Cancer With Triple Negative Breast Cancer

Single progesterone receptor positive (PgR+), especially in form of ER-/PgR+/HER2-, is a nonnegligible phenomenon. Little is known about the characteristics and the role of single PgR positive in this phenotype. Therefore, we explore the significance of single PgR positivity by comparing ER-/PgR+/HER2- breast cancers with triple negative breast cancers (TNBCs).Three thousand nine hundred sixty-six cases of primary invasive breast carcinoma operated consecutively from January 2005 to May 2008 in Cancer Hospital, Chinese Academy of Medical Sciences were examined. Two hundred forty (6%) cases were identified as ER-/PgR+/HER2- breast cancers and 348 (8.8%) cases as TNBCs. Clinicopathological characteristics and survivals were analyzed respectively and then compared between 2 subtypes.Compared with patients with TNBCs, ER-/PgR+/HER2- tumor tended to have lower tumor grade (Grade 3: 45.7% vs. 37.5%, P = 0.051) and smaller tumor size (P = 0.036). However, no differences were found between ER-/PgR+/HER2- and TNBC patients in relapse-free survival (RFS) and OS. The 5-year RFS rates were 80.7% and 77.4%, respectively (P = 0.330) and the 5-year OS rates were 88.0% and 85.2%, respectively (P = 0.290). ER-/PgR+/HER2- patients receiving adjuvant endocrine treatment had better RFS (P = 0.016) and overall survival (OS) (P < 0.0001) than patients receiving no endocrine therapy.This exclusive analysis of patients with ER-/PgR+/HER2- breast cancers showed that this subtype exhibited an aggressive behavior as TNBC, suggesting that it should also be regarded as biologically distinctive group and single PgR positive itself is not a good prognostic factor. However, adjuvant endocrine therapy could still benefit this group of patients. Further investigations should be done to elucidate the underlying mechanism.

Fine-needle aspirate CYFRA 21-1, an innovative new marker for diagnosis of axillary lymph node metastasis in breast cancer patients

To compare the value of cytokeratin fragment 21-1 (CYFRA 21-1) concentration in the fine-needle biopsy aspirates (fine needle aspirate [FNA] CYFRA 21-1) with cytopathology of fine-needle aspiration biopsy (FNA cytology) and to assess whether CYFRA 21-1 concentrations from ultrasound-guided fine-needle aspiration biopsy (US-FNAB) specimens (FNA CYFRA 21-1) is not inferior to FNA cytology in the diagnosis of axillary lymph node (ALN) metastasis of breast cancer patients.This study received institutional review board approval, and written informed consent was obtained from all patients. US-FNAB was performed in 373 ALNs from 358 patients with invasive breast cancer. Concentrations of CYFRA 21-1 were measured from washouts of the syringe used during US-FNAB (FNA CYFRA 21-1), and ALN metastasis was determined using a cutoff value of 1.93 ng/mL. FNA cytology, intraoperative sentinel lymph node biopsy, and surgical pathology results were reviewed and analyzed. The noninferiority margin for the difference in accuracies between FNA cytology and FNA CYFRA 21-1 was set as 5%.Among 373 ALNs, 136 (36.5%) were benign, and 237 (63.5%) were metastatic. The mean FNA CYFRA 21-1 was significantly higher in metastatic ALNs compared to that in benign ALNs (P = 0.001). For the diagnosis of ALN metastasis, the sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of FNA CYFRA 21-1 (cutoff value 1.93 ng/mL) were not significantly different from those of FNA cytology (P > 0.05). FNA CYFRA 21-1 reached statistical noninferiority to FNA cytology in terms of diagnostic accuracy for ALN metastasis. Of the 20 ALNs (8 metastasis, 12 benign) that showed insufficient results on FNA cytology, FNA CYFRA 21-1 accurately diagnosed 15 ALNs (4 metastasis, 11 benign).The diagnostic performance of FNA CYFRA 21-1 is comparable with that of FNA cytology for breast cancer ALN metastasis. Our results indicate that FNA CYFRA 21-1, using an US-FNAB specimen, can be a useful method equal to FNA cytology in terms of diagnostic accuracy.

Computer-aided detection (CAD) system for breast MRI in assessment of local tumor extent, nodal status, and multifocality of invasive breast cancers: preliminary study

Background

We aimed to investigate the efficacy of computer-aided detection (CAD) for MRI in the assessment of tumor extent, lymph node status, and multifocality in invasive breast cancers in comparison with other breast imaging modalities.

Methods

Two radiologists measured the maximum tumor size, as well as, analyzed lymph node status and multifocality in 86 patients with invasive breast cancers using mammography, ultrasound, CT, MRI with and without CAD, and 18-fludeoxyglucose positron emission tomography (FDG-PET). The assessed data were compared with pathology.

Results

For tumor extent, there were no significant differences between pathological size and measured size using mammography, ultrasound, CT, or MRI with and without CAD (P > 0.05). For evaluation of lymph node status, ultrasound had the best kappa coefficients (0.522) for agreement between imaging and pathology, and diagnostic performance with 92.1% specificity and 90.0% positive predictive value. For multifocality, MRI with CAD had the highest area under the receiver operating characteristic curve (AUC = 0.888).

Conclusions

CAD for MRI is feasible to assess tumor extent and multifocality in invasive breast cancer patients. However, CAD is not effective in evaluation of nodal status.

Diffusion weighted images of metastatic as compared with nonmetastatic axillary lymph nodes in patients with newly diagnosed breast cancer

BACKGROUND

To investigate the ability of diffusion weighted images (DWI) to differentiate between metastatic and nonmetastatic axillary lymph nodes (LNs) in patients with newly diagnosed breast cancer.

METHODS

From January 2010 to February 2012, DWI was performed at b values of 0 and 800 for 16 metastatic LNs from 16 patients with breast cancer, and 20 nonmetastatic LNs from 20 women without breast cancer. The metastatic LNs were proven by ultrasound (US) guided core biopsy and the same LNs were identified on MRI by comparing the US images with MR images. Nonmetastatic LNs were verified by the stability in size and shape for at least 2 years on MRI. The apparent diffusion coefficient (ADC) value of the metastatic and nonmetastatic axillary LNs was compared. Receiver-operating-characteristics (ROC) analysis was performed to evaluate the diagnostic performance of the ADC value in differentiating between metastatic and nonmetastatic axillary LNs.

RESULTS

The mean ADC value was 0.746 × 10(-3) for metastatic LNs and 1.033 × 10(-3) for nonmetastatic LNs (P < 0.001). The area under the ROC curve was 0.884. The sensitivity and specificity for differentiating metastatic from nonmetastatic axillary LNs using a cutoff ADC value of 0.852 were 85% and 81%, respectively.

CONCLUSION

There is a statistically significant difference between the ADC values of pathologically proven metastatic LNs and nonmetastatic LNs. DWI and ADC values are a useful tool for differentiating metastatic from nonmetastatic axillary LNs.

HER2 expression in fine needle aspirates of lymph nodes detected by preoperative axillary ultrasound in breast cancer patients

The purpose of this study was to assess the usefulness of HER2 levels in ultrasonographically guided fine-needle aspiration biopsy (US-FNA) aspirates of axillary lymph nodes (ALNs) in the determination of lymph node metastasis or the characterization of primary breast cancer, and to correlate the HER2 levels in US-FNA aspirates (FNA-HER2s) of metastatic ALNs with the HER2 statuses of corresponding primary breast cancers. An institutional review board approved the study. Between January and October 2010, 164 patients with 167 ALNs examined by US-FNA were included. FNA-HER2s of ALNs were measured by chemiluminescence immunoassay, and they were correlated with cytologic/final diagnoses. Receiver operating characteristics (ROC) curve analysis was performed to evaluate the diagnostic ability to differentiate benign and metastatic ALNs. Additionally, FNA-HER2s of metastatic ALNs were correlated with HER2 status and other clinicopathologic variables of the primary breast cancers. Among the 167 ALNs, 138 were metastatic and 29 were benign. The mean FNA-HER2 (6.3 ng/ml) of metastatic ALNs was higher than that of benign ALNs. All 29 benign ALNs showed no measurable value of FNA-HER2 (0.0 ng/ml). The area under the ROC curves of FNA-HER2 of ALNs was 0.679 for the diagnosis of ALN metastasis. The FNA-HER2 statuses of 108 metastatic ALNs (79.4%) were concordant with the HER2 statuses of the corresponding primary breast cancers. In a subgroup analysis of HER2-positive cancers with ALN metastasis, distant metastasis was significantly associated with FNA-HER2-negativity of metastatic ALNs (P = 0.04). Although FNA-HER2 of ALNs did not improve the diagnostic performance of FNA cytology in preoperative diagnosis of ALN metastasis of overall patients, FNA-HER2-positive metastatic ALNs were significantly associated with HER2-positivity of primary breast cancers. Additionally, FNA-HER2 analysis of ALN may help to develop more personalized treatment protocol for breast cancer patients by determining the concordance or discordance of HER2 status between primary cancers and metastatic ALNs.

The clinical value of hybrid sentinel lymphoscintigraphy to predict metastatic sentinel lymph nodes in breast cancer

PURPOSE

Hybrid imaging techniques can provide functional and anatomical information about sentinel lymph nodes in breast cancer. Our aim in this study was to evaluate which imaging parameters on hybrid sentinel lymphoscintigraphy predicted metastatic involvement of sentinel lymph nodes (SLNs) in patients with breast cancer.

METHODS

Among 56 patients who underwent conventional sentinel lymphoscintigraphy, 45 patients (age, 53.1 ± 9.5 years) underwent hybrid sentinel lymphoscintigraphy using a single-photon emission computed tomography (SPECT)/computed tomography (CT) gamma camera. On hybrid SPECT/CT images, we compared the shape and size (long-to-short axis [L/S] ratio) of the SLN, and SLN/periareolar injection site (S/P) count ratio between metastatic and non-metastatic SLNs. Metastatic involvement of sentinel lymph nodes was confirmed by pathological biopsy.

RESULTS

Pathological biopsy revealed that 21 patients (46.7 %) had metastatic SLNs, while 24 (53.3 %) had non-metastatic SLNs. In the 21 patients with metastatic SLNs, the SLN was mostly round (57.1 %) or had an eccentric cortical rim (38.1 %). Of 24 patients with non-metastatic SLNs, 13 patients (54.1 %) had an SLN with a C-shape rim or eccentric cortex. L/S ratio was 2.04 for metastatic SLNs and 2.38 for non-metastatic SLNs. Seven (33 %) patients had T1 primary tumors and 14 (66 %) had T2 primary tumors in the metastatic SLN group. In contrast, 18 (75 %) patients had T1 primary tumors and six (25 %) had T2 tumors in the non-metastatic SLN group. S/P count ratio was significantly lower in the metastatic SLN group than the non-metastatic SLN group for those patients with a T1 primary tumor (p = 0.007).

CONCLUSIONS

Hybrid SPECT/CT offers the physiologic data of SPECT together with the anatomic data of CT in a single image. This hybrid imaging improved the anatomic localization of SLNs in breast cancer patients and predicted the metastatic involvement of SLNs in the subgroup of breast cancer patients with T1 primary tumors.

Tree-based model for thyroid cancer prognostication

BACKGROUND

Death is uncommon in thyroid cancer patients, and the factors important in predicting survival remain inadequately studied. The objective of this study was to assess prognostic effects of patient, tumor, and treatment factors and to determine prognostic groups for thyroid cancer survival.

METHODS

Using data from the Surveillance, Epidemiology, and End Results Program (SEER), we evaluated overall and disease-specific survival (DSS) in 43 392 well-differentiated thyroid cancer patients diagnosed from 1998 through 2005. Multivariable analyses were performed using Cox proportional hazards regression, survival trees, and random survival forest. Similar analyses were performed using National Cancer Data Base data, with overall survival (OS) evaluated in 131 484 thyroid cancer patients diagnosed from 1998 through 2005. Relative importance of factors important to survival was assessed based on the random survival forest analyses.

RESULTS

Using survival tree analyses, we identified 4 distinct prognostic groups based on DSS (P < .0001). The 5-year DSS of these prognostic groups was 100%, 98%, 91%, 64%, whereas the 10-year survival was 100%, 96%, 85%, and 50%. Based on random survival forest analyses, the most important factors for DSS were SEER stage and age at diagnosis. For OS, important prognostic factors were similar, except age at diagnosis demonstrated marked importance relative to SEER stage. Similar results for OS were found using National Cancer Data Base data.

CONCLUSION

This study identifies distinct prognostic groups for thyroid cancer and illustrates the importance of patient age to both disease-specific and OS. These findings have implications for patient education and thyroid cancer treatment.

Lymphoscintigraphy with peritumoral injection versus lymphoscintigraphy with subdermal periareolar injection of technetium-labeled human albumin to identify sentinel lymph nodes in breast cancer patients

BACKGROUND

Preoperative lymphoscintigraphy is without doubt a valid method for the detection of the sentinel lymph node (SLN). There has been considerable debate regarding the optimal site for the introduction of the tracer; various sites include peritumoral (PT), periareolar (PA), subdermal, and intradermal injection.

PURPOSE

To evaluate retrospectively the lymphoscintigraphic identification rate of peritumoral (PT) injection versus subdermal periareolar (PA) injection in the detection of SLNs in breast cancer.

MATERIAL AND METHODS

Between October 2002 and December 2011, a cohort of 906 consecutive patients with biopsy proven breast cancer underwent 914 SLN biopsy procedures. A total of 339 procedures (Group A) were performed using a PT deep injection of radiotracer while 575 procedures (Group B) adopted a subdermal PA injection of radiotracer towards the upper outer quadrant, regardless of the site of the carcinoma. All the patients underwent synchronous excision of the breast cancer and SLN biopsy.

RESULTS

SLNs were identified in the lymphoscintigram in 308/339 cases (90.85%) of Group A (PT injection) and in 537/575 cases (93.39%) of Group B (PA injection). Furthermore, in 2/339 patients (0.58%) of Group A, internal mammary lymph nodes were found at lymphoscintigraphy, whereas no internal mammary sentinel nodes were found in the Group B patients. The intraoperative identification rate of axillary SLNs was 99.41% (337 of 339) in the Group A patients and 99.65% (573 of 575) in the Group B patients. There was no significant difference in the two groups between the incidence of the number of SLNs detected and the incidence of identification of positive SLNs.

CONCLUSION

PT versus PA injection of radiotracer showed comparable success rates for axillary SLN identification, and can be considered a rapid and reliable method.

Accuracy of axillary lymph node staging in breast cancer patients: an observer-performance study comparison of MRI and ultrasound

PURPOSE

To compare magnetic resonance imaging (MRI) and ultrasound (US) for axillary lymph node (LN) staging in breast cancer patients in an observer-performance study.

MATERIALS AND METHODS

An observer-performance study was conducted with five breast radiologists reviewing 50 consecutive patients of newly diagnosed invasive breast cancer with the use of ipsilateral axillary MRI and US. LN status was pathologically proved in all patients. Each observer reviewed the images in two separate sessions: one for MRI and the other for US. Observers were asked to indicate their confidence of the presence of at least one ipsilateral metastatic LN on a quasi-continuous rating scale and whether they recommend percutaneous biopsy preoperatively. Receiver operating characteristic (ROC) analysis and area under the ROC curve were used to characterize diagnostic performance. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated from whether observers recommended biopsy.

RESULTS

There were no statistically significant differences in each observer's performance between MRI and US, or in the performance of all observers as a group, in terms of ROC analysis. There were no statistically significant differences in sensitivity, specificity, PPV, or NPV between MRI and US, but there were statistically significant improvements in specificity and PPV from either MRI or US alone to MRI and US combined.

CONCLUSIONS

Observer performance on MRI and US are comparable for axillary LN staging. When US and MRI are concordant for positive findings, higher specificity and PPV can be obtained.

False negative results of preoperative axillary ultrasound in patients with invasive breast cancer: correlations with clinicopathologic findings

The purpose of this study was to investigate false-negative results of preoperative axillary ultrasound (US) and to evaluate clinicopathologic factors related to false-negative results in patients with invasive breast cancer. Four-hundred eighty-two patients with 483 invasive breast cancers who had no suspicious findings on preoperative axillary US were included in this study. All patients underwent sentinel lymph node biopsy (SLNB) or axillary lymph node dissection. False-negative and true-negative results were compared in terms of age, body mass index (BMI), T-stage, pathologic N-stage and final diagnosis of breast cancer. Statistical analyses were performed using the chi-square or Fisher's exact tests. Of the 483 axillae with negative results on axillary US, 93 axillae of 93 patients showed false-negative results and the negative predictive value of axillary US was 80.8% (390/483). Seventy-five axillae (15.5%, 75/483) had N1 and 18 axillae (3.7%, 18/483) had N2 or N3 disease. Eighteen false-negative results with N2/N3 disease showed a significantly higher T stage (T2/T3) than those with N1 disease. As the T-stage increased, false-negative results were found more often on preoperative axillary US (p < 0.05). Age, BMI and final diagnosis of primary breast cancer were not associated with false-negative results on preoperative axillary US. Preoperative axillary US alone is insufficiently specific to obviate the need for SLNB because of the substantial number of false-negative results in patients with invasive breast cancer, although preoperative axillary US alone may exclude most cases of N2 and N3 disease.