Deep learning radiomics for prediction of axillary lymph node metastasis in patients with clinical stage T1-2 breast cancer

Prediction of axillary lymph node metastasis in T1 breast cancer using diffuse optical tomography, strain elastography and molecular markers

Background

Early-stage breast cancer (BC) presents a certain risk of axillary lymph node (ALN) metastasis (ALNM), leading to different individualized treatment. Preoperative non-invasive prediction to determine ALN status is of great significance for avoiding ineffective axillary surgery. Tumor total hemoglobin concentration (TTHC), strain ratio (SR), and Ki-67 expression are associated with ALNM in BC, but few studies have focused on T1 BC. This study aimed to explore the usefulness of these factors individually and in combination for the preoperative prediction of ALNM in T1 BC.

Methods

This was a cross-sectional study. A total of 122 patients with T1 BC were enrolled. TTHC and SR were assessed preoperatively using diffuse optical tomography and strain elastography, respectively. All patients were pathologically evaluated to determine ALN status. Univariate analysis and logistic regression trend test were performed to identify independent predictors. A combined model of imaging-pathological parameters for ALNM was developed.

Results

Histopathological analysis indicated that 56 patients (45.9%) exhibited ALNM. The fully adjusted model demonstrated a significant trend correlating increased TTHC (P<0.01 for trend), SR (P<0.001 for trend), and Ki-67 expression (P=0.004 for trend) with ALNM. The area under the receiver operating characteristic curves (AUCs) of TTHC, SR, and Ki-67 expression for predicting ALNM were 0.707 [95% confidence interval (CI): 0.61-0.80], 0.718 (95% CI: 0.63-0.81) and 0.642 (95% CI: 0.56-0.73), respectively. The integration of imaging parameters (TTHC and SR) and Ki-67 expression yielded superior predictive performance for ALN status compared to each parameter individually, as evidenced by an AUC of 0.837 (95% CI: 0.76-0.91). The Hosmer-Lemeshow test P value was 0.880, demonstrating good calibration. In the subgroup analysis, the model exhibited positive predictive capabilities (AUC >0.75) across various subgroups, including molecular subtype, pathological type, and grade. Notably, the predictive performance was particularly enhanced in patients with invasive lobular carcinomas and triple-negative BC, with AUC values exceeding 0.90 for both subgroups.

Conclusions

The study found that TTHC ≥185.75 µmol/L, SR ≥3.93 and Ki-67 expression ≥20% were strongly associated with ALNM in patients with T1 BC. The combined imaging-pathological model might provide a convenient preoperative method for predicting ALN status, which can assist clinicians in individualizing management for patients with early-stage BC.

Artificial intelligence performance in ultrasound-based lymph node diagnosis: a systematic review and meta-analysis

BACKGROUND AND OBJECTIVES

Accurate classification of lymphadenopathy is essential for determining the pathological nature of lymph nodes (LNs), which plays a crucial role in treatment selection. The biopsy method is invasive and carries the risk of sampling failure, while the utilization of non-invasive approaches such as ultrasound can minimize the probability of iatrogenic injury and infection. With the advancement of artificial intelligence (AI) and machine learning, the diagnostic efficiency of LNs is further enhanced. This study evaluates the performance of ultrasound-based AI applications in the classification of benign and malignant LNs.

METHODS

The literature research was conducted using the PubMed, EMBASE, and Cochrane Library databases as of June 2024. The quality of the included studies was evaluated using the QUADAS-2 tool. The pooled sensitivity, specificity, and diagnostic odds ratio (DOR) were calculated to assess the diagnostic efficacy of ultrasound-based AI in classifying benign and malignant LNs. Subgroup analyses were also conducted to identify potential sources of heterogeneity.

RESULTS

A total of 1,355 studies were identified and reviewed. Among these studies, 19 studies met the inclusion criteria, and 2,354 cases were included in the analysis. The pooled sensitivity, specificity, and DOR of ultrasound-based machine learning in classifying benign and malignant LNs were 0.836 (95% CI [0.805, 0.863]), 0.850 (95% CI [0.805, 0.886]), and 33.331 (95% CI [22.873, 48.57]), respectively, indicating no publication bias (p = 0.12). Subgroup analyses may suggest that the location of lymph nodes, validation methods, and type of primary tumor are the sources of heterogeneity.

CONCLUSION

AI can accurately differentiate benign from malignant LNs. Given the widespread use of ultrasonography in diagnosing malignant LNs in cancer patients, there is significant potential for integrating AI-based decision support systems into clinical practice to enhance the diagnostic accuracy.

Prediction of axillary nodal burden using preoperative magnetic resonance imaging scoring in patients with clinically node-negative breast cancer: a retrospective cohort study

Background

Axillary lymph node metastasis (ALNM) is a significant predictor of overall patient survival; thus, precise evaluation of ALNM is essential for staging breast cancer, informing multimodal treatment strategies, and ensuring optimal patient care. This study aimed to establish a magnetic resonance imaging (MRI) scoring system for predicting extensive axillary nodal metastasis in patients with clinically node-negative breast cancer derived from preoperative breast and axillary MRI.

Methods

This study included 226 patients with clinically node-negative breast cancer who underwent preoperative breast and axillary MRI between January 1, 2010 and December 31, 2020 at King Chulalongkorn Memorial Hospital. Their clinical, radiological, and pathological features were retrospectively reviewed. MRI characteristics of breast tumors and axillary lymph nodes (LNs) were assessed. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and Cohen's Kappa coefficient of the scoring system were evaluated. The receiver operating characteristic curve was used to determine the cutoff value for the MRI scoring system to differentiate extensive ALNM from nonextensive ALNM.

Results

Of the 226 patients, 144 had cancer-free axilla, 51 had 1-2 positive metastatic LNs, and 31 had ≥3 positive metastatic LNs. Moreover, only 60 could be evaluated for the apparent diffusion coefficient (ADC) value of LNs because of size limitations. The cutoff value for the MRI scoring system with ADC was 14 (NPV =87.1% with moderately acceptable discrimination), and the cutoff value without ADC was 8 (sensitivity =77.4%; specificity =81%; PPV =39.3%; NPV =95.8% with moderately acceptable discrimination).

Conclusions

The MRI scoring system using breast and axillary LN characteristics from preoperative MRI may help predict extensive ALNM and aid axillary nodal treatment selection.