Imaging evaluation of lymphadenopathy and patterns of lymph node spread in head and neck cancer

Automated Segmentation of Lymph Nodes on Neck CT Scans Using Deep Learning

Early and accurate detection of cervical lymph nodes is essential for the optimal management and staging of patients with head and neck malignancies. Pilot studies have demonstrated the potential for radiomic and artificial intelligence (AI) approaches in increasing diagnostic accuracy for the detection and classification of lymph nodes, but implementation of many of these approaches in real-world clinical settings would necessitate an automated lymph node segmentation pipeline as a first step. In this study, we aim to develop a non-invasive deep learning (DL) algorithm for detecting and automatically segmenting cervical lymph nodes in 25,119 CT slices from 221 normal neck contrast-enhanced CT scans from patients without head and neck cancer. We focused on the most challenging task of segmentation of small lymph nodes, evaluated multiple architectures, and employed U-Net and our adapted spatial context network to detect and segment small lymph nodes measuring 5-10 mm. The developed algorithm achieved a Dice score of 0.8084, indicating its effectiveness in detecting and segmenting cervical lymph nodes despite their small size. A segmentation framework successful in this task could represent an essential initial block for future algorithms aiming to evaluate small objects such as lymph nodes in different body parts, including small lymph nodes looking normal to the naked human eye but harboring early nodal metastases.

Development and validation of a nomogram to predict cervical lymph node metastasis in head and neck squamous cell carcinoma

Background

Head and neck cancers are a heterogeneous, aggressive, and genetically complex collection of malignancies of the oral cavity, nasopharynx, oropharynx, hypopharynx, larynx, paranasal sinuses and salivary glands, which are difficult to treat. Regional lymph nodes metastasis is a significant poor prognosis factor for head and neck squamous cell carcinoma. Metastasis to the regional lymph nodes reduces the 5-year survival rate by 50% compared with that of patients with early-stage disease. Accurate evaluation of cervical lymph node is a vital component in the overall treatment plan for patients with squamous cell carcinoma of the head and neck. However, current models are struggle to accurately to predict cervical lymph node metastasis. Here, we analyzed the clinical, imaging, and pathological data of 272 patients with HNSCC confirmed by postoperative pathology and sought to develop and validate a nomogram for prediction of lymph node metastasis in patients with head and neck squamous cell carcinoma.

Methods

We retrospectively analyzed the clinical, imaging, and pathological data of 272 patients with head and neck squamous cell carcinoma (HNSCC) confirmed by postoperative pathology at the Affiliated Hospital of Qingdao University from June 2017 to June 2021. Patients were randomly divided into the training and validation cohorts in a 3:1 ratio, and after screening risk factors by logistic regression, nomogram was developed for predicting lymph nodes metastasis, then the prediction model was verified by C-index, area under curve (AUC), and calibration curve.

Results

Of the 272 patients, seven variables were screened to establish the predictive model, including the differentiation degree of the tumor [95% confidence interval(CI):1.224~6.735, P=0.015], long-to-short axis ratio of the lymph nodes (95%CI: 0.019~0.217, P<0.001), uneven/circular enhancement (95%CI: 1.476~16.715, P=0.010), aggregation of lymph nodes (95%CI:1.373~10.849, P=0.010), inhomogeneous echo (95%CI: 1.337~23.389, P=0.018), unclear/absent medulla of lymph nodes (95%CI: 2.514~43.989, P=0.001), and rich blood flow (95%CI: 1.952~85.632, P=0.008). The C-index was 0.910, areas under the curve of training cohort and verification cohort were 0.953 and 0.938 respectively, indicating the discriminative ability of this nomogram. The calibration curve showed a favorable compliance between the prediction of the model and actual observations. The clinical decision curve showed this model is clinically useful and had better discriminative ability between 0.25 and 0.9 for the probability of cervical LNs metastasis.

Conclusions

We established a good prediction model for cervical lymph node metastasis in head and neck squamous cell carcinoma patients which can provide reference value and auxiliary diagnosis for clinicians in making neck management decisions of HNSCC patients.

A deep learning and radiomics fusion model based on contrast-enhanced computer tomography improves preoperative identification of cervical lymph node metastasis of oral squamous cell carcinoma

OBJECTIVES

In this study, we constructed and validated models based on deep learning and radiomics to facilitate preoperative diagnosis of cervical lymph node metastasis (LNM) using contrast-enhanced computed tomography (CECT).

MATERIALS AND METHODS

CECT scans of 100 patients with OSCC (217 metastatic and 1973 non-metastatic cervical lymph nodes: development set, 76 patients; internally independent test set, 24 patients) who received treatment at the Peking University School and Hospital of Stomatology between 2012 and 2016 were retrospectively collected. Clinical diagnoses and pathological findings were used to establish the gold standard for metastatic cervical LNs. A reader study with two clinicians was also performed to evaluate the lymph node status in the test set. The performance of the proposed models and the clinicians was evaluated and compared by measuring using the area under the receiver operating characteristic curve (AUC), accuracy (ACC), sensitivity (SEN), and specificity (SPE).

RESULTS

A fusion model combining deep learning with radiomics showed the best performance (ACC, 89.2%; SEN, 92.0%; SPE, 88.9%; and AUC, 0.950 [95% confidence interval: 0.908-0.993, P < 0.001]) in the test set. In comparison with the clinicians, the fusion model showed higher sensitivity (92.0 vs. 72.0% and 60.0%) but lower specificity (88.9 vs. 97.5% and 98.8%).

CONCLUSION

A fusion model combining radiomics and deep learning approaches outperformed other single-technique models and showed great potential to accurately predict cervical LNM in patients with OSCC.

CLINICAL RELEVANCE

The fusion model can complement the preoperative identification of LNM of OSCC performed by the clinicians.

Ultrasound-guided hookwire localization of non palpable cervical lymphadenopathy: A case-control study of operative time

OBJECTIVE

We aimed at evaluating the impact of ultrasound-guided (US) hookwire localization of nonpalpable cervical lymphadenopathy on operating time.

DESIGN AND METHODS

Retrospective case control study (January 2017 and May 2021) of 26 patients with lateral nonpalpable cervical lymphadenopathy undergoing surgery with (H+) and without (H-) per operative US-guided hook-wire localization. Operative time (general anesthesiology onset, hookwire placement, end of surgery) and surgery-related adverse events data were collected.

RESULTS

Mean operative time was significantly shorter in H+ group versus H- group (26 ± 16 min vs. 43 ± 22 min) (p = 0.02). Histopathological diagnosis accuracy was 100% versus 94% (H+ vs. H-, p = 0.1). No significant between group difference in surgery-related adverse events was reported (wound healing, p = 0.162; hematomas, p = 0.498; neoplasms removal failure, p = 1).

CONCLUSION

US-guided hookwire localization of lateral nonpalpable cervical lymphadenopathy allowed a significant reduction in operative time, comparable histopathological diagnosis accuracy and adverse events compared with H-.

Efficacy of texture analysis of ultrasonographic images in the differentiation of metastatic and non-metastatic cervical lymph nodes in patients with squamous cell carcinoma of the tongue

OBJECTIVE

We investigated the efficacy of using texture analysis of ultrasonographic images of the cervical lymph nodes of patients with squamous cell carcinoma of the tongue to differentiate between metastatic and non-metastatic lymph nodes.

STUDY DESIGN

We analyzed 32 metastatic and 28 non-metastatic lymph nodes diagnosed by histopathologic examination on presurgical US images. Using the LIFEx texture analysis program, we extracted 36 texture features from the images and calculated the statistical significance of differences in texture features between metastatic and non-metastatic lymph nodes using the t test. To assess the diagnostic ability of the significantly different texture features to discriminate between metastatic and non-metastatic nodes, we performed receiver operating characteristic curve analysis and calculated the area under the curve. We set the cutoff points that maximized the sensitivity and specificity for each curve according to the Youden J statistic.

RESULTS

We found that 20 texture features significantly differed between metastatic and non-metastatic lymph nodes. Among them, only the gray-level run length matrix feature of run length non-uniformity and the gray-level zone length matrix features of gray-level non-uniformity and zone length non-uniformity showed an excellent ability to discriminate between metastatic and non-metastatic lymph nodes as indicated by the area under the curve and the sum of sensitivity and specificity.

CONCLUSIONS

Analysis of the texture features of run length non-uniformity, gray-level non-uniformity, and zone length non-uniformity values allows for differentiation between metastatic and non-metastatic lymph nodes, with the use of gray-level non-uniformity appearing to be the best means of predicting metastatic lymph nodes.

Clinical application of MR-Linac in tumor radiotherapy: a systematic review

Recent years have seen both a fresh knowledge of cancer and impressive advancements in its treatment. However, the clinical treatment paradigm of cancer is still difficult to implement in the twenty-first century due to the rise in its prevalence. Radiotherapy (RT) is a crucial component of cancer treatment that is helpful for almost all cancer types. The accuracy of RT dosage delivery is increasing as a result of the quick development of computer and imaging technology. The use of image-guided radiation (IGRT) has improved cancer outcomes and decreased toxicity. Online adaptive radiotherapy will be made possible by magnetic resonance imaging-guided radiotherapy (MRgRT) using a magnetic resonance linear accelerator (MR-Linac), which will enhance the visibility of malignancies. This review's objectives are to examine the benefits of MR-Linac as a treatment approach from the perspective of various cancer patients' prognoses and to suggest prospective development areas for additional study.

[Ultrasound of the neck]

Ultrasound examination of the neck organs enables an assessment that in many cases is superior to that of magnetic resonance imaging and computed tomography. Ultrasound is therefore not only a first line or point of care imaging modality but can provide imaging for the concluding diagnosis in cases. Because of the good sonographic accessibility of the majority of the structures of the neck, many technical advances, in particular high-resolution ultrasound and signal post-processing have a major influence on the possibilities of ultrasound. Lymph nodes and salivary glands are the main focus in clinical applications, although other diseases and swellings of the neck can also be clarified with ultrasound. Special applications are ultrasound-guided interventions, e.g., biopsies or the sonographic assessment of peripheral nerves. As in any imaging modality, a comprehensive clinical knowledge is necessary for the diagnostic assessment. Because of constant assessment and thus continuous modification of the examination, ultrasound examinations may only be performed adequately with the appropriate clinical knowledge.

Preliminary study on CT contrast-enhanced radiomics for predicting central cervical lymph node status in patients with thyroid nodules

Objective

To explore the feasibility of using a contrast-enhanced CT image-based radiomics model to predict central cervical lymph node status in patients with thyroid nodules.

Methods

Pretreatment clinical and CT imaging data from 271 patients with surgically diagnosed and treated thyroid nodules were retrospectively analyzed. According to the pathological features of the thyroid nodules and central lymph nodes, the patients were divided into three groups: group 1: papillary thyroid carcinoma (PTC) metastatic lymph node group; group 2: PTC nonmetastatic lymph node group; and group 3: benign thyroid nodule reactive lymph node group. Radiomics models were constructed to compare the three groups by pairwise classification (model 1: group 1 vs group 3; model 2: group 1 vs group 2; model 3: group 2 vs group 3; and model 4: group 1 vs groups (2 + 3)). The feature parameters with good generalizability and clinical risk factors were screened. A nomogram was constructed by combining the radiomics features and clinical risk factors. Receiver operating characteristic (ROC) curve, calibration curve and decision curve analysis (DCA) were performed to assess the diagnostic and clinical value of the nomogram.

Results

For radiomics models 1, 2, and 3, the areas under the curve (AUCs) in the training group were 0.97, 0.96, and 0.93, respectively. The following independent clinical risk factors were identified: model 1, arterial phase CT values; model 2, sex and arterial phase CT values; model 3: none. The AUCs for the nomograms of models 1 and 2 in the training group were 0.98 and 0.97, respectively, and those in the test group were 0.95 and 0.87, respectively. The AUCs of the model 4 nomogram in the training and test groups were 0.96 and 0.94, respectively. Calibration curve analysis and DCA revealed the high clinical value of the nomograms of models 1, 2 and 4.

Conclusion

The nomograms based on contrast-enhanced CT images had good predictive efficacy in classifying benign and malignant central cervical lymph nodes of thyroid nodule patients.

Deep learning assisted contrast-enhanced CT-based diagnosis of cervical lymph node metastasis of oral cancer: a retrospective study of 1466 cases

OBJECTIVES

Lymph node (LN) metastasis is a common cause of recurrence in oral cancer; however, the accuracy of distinguishing positive and negative LNs is not ideal. Here, we aimed to develop a deep learning model that can identify, locate, and distinguish LNs in contrast-enhanced CT (CECT) images with a higher accuracy.

METHODS

The preoperative CECT images and corresponding postoperative pathological diagnoses of 1466 patients with oral cancer from our hospital were retrospectively collected. In stage I, full-layer images (five common anatomical structures) were labeled; in stage II, negative and positive LNs were separately labeled. The stage I model was innovatively employed for stage II training to improve accuracy with the idea of transfer learning (TL). The Mask R-CNN instance segmentation framework was selected for model construction and training. The accuracy of the model was compared with that of human observers.

RESULTS

A total of 5412 images and 5601 images were labeled in stage I and II, respectively. The stage I model achieved an excellent segmentation effect in the test set (AP₅₀-0.7249). The positive LN accuracy of the stage II TL model was similar to that of the radiologist and much higher than that of the surgeons and students (0.7042 vs. 0.7647 (p = 0.243), 0.4216 (p < 0.001), and 0.3629 (p < 0.001)). The clinical accuracy of the model was highest (0.8509 vs. 0.8000, 0.5500, 0.4500, and 0.6658 of the Radiology Department).

CONCLUSIONS

The model was constructed using a deep neural network and had high accuracy in LN localization and metastasis discrimination, which could contribute to accurate diagnosis and customized treatment planning.

KEY POINTS

• Lymph node metastasis is not well recognized with modern medical imaging tools. • Transfer learning can improve the accuracy of deep learning model prediction. • Deep learning can aid the accurate identification of lymph node metastasis.

Magnetic Resonance-Guided Radiation Therapy for Head and Neck Cancers

Despite the significant evolution of radiation therapy (RT) techniques in recent years, many patients with head and neck cancer still experience significant toxicities during and after treatments. The increased soft tissue contrast and functional sequences of magnetic resonance imaging (MRI) are particularly attractive in head and neck cancer and have led to the increasing development of magnetic resonance-guided RT (MRgRT). This approach refers to the inclusion of the additional information acquired from a diagnostic or planning MRI in radiation treatment planning, and now extends to online high-quality daily imaging generated by the recently developed MR-Linac. MRgRT holds numerous potentials, including enhanced baseline and planning evaluations, anatomical and functional treatment adaptation, potential for hypofractionation, and multiparametric assessment of response. This article offers a structured review of the current literature on these established and upcoming roles of MRI for patients with head and neck cancer undergoing RT.

Inter-rater concordance and operating definitions of radiologic nodal feature assessment in human papillomavirus-positive oropharyngeal carcinoma

BACKGROUND AND PURPOSE

This study aims to evaluate the reliability of radiologic nodal feature assessment in clinical node-positive human papillomavirus-positive oropharyngeal carcinoma.

MATERIALS AND METHODS

Baseline CTs or MRIs of clinical node-positive human papillomavirus-positive oropharyngeal carcinoma diagnosed between 2012 and 2015 were reviewed independently by two neuroradiologists for seven nodal features: radiologic nodal involvement, cystic change, presence of necrosis, clustering, conglomeration, coalescence, and extranodal extension. Consensus operating definitions were derived after discussion. The features were re-reviewed in a randomly selected cohort. Levels of certainty (probability of presence: <25%, ∼50%, ∼75%, and >90%) were recorded. Interrater concordance was calculated using Cohen's kappa coefficient.

RESULTS

A total of 413 patients (826 necks) were eligible. At initial review, the inter-rater kappa values for: radiologic nodal involvement, cystic change, necrosis, clustering, conglomeration, coalescence, and extranodal extension were 0.92, 0.64, 0.48, 0.32, 0.32, 0.62, and 0.56, respectively. A re-review of 94 randomly selected cases (188 necks) after consolidation of operating definitions for nodal features showed that the inter-rater kappa values of these features were 0.83, 0.62, 0.58, 0.32, 0.18, 0.68, and 0.74 when considering ≥50% certainty as positive, and improved to 0.94, 0.66, 0.59, 0.33, 0.19, 0.76, and 0.86 when considering ≥75% certainty as positive.

CONCLUSION

Clearly defined nomenclature results in improved interrater reliability when assessing radiologic nodal features, especially for coalescent adenopathy and extranodal extension. Higher levels of certainty are associated with higher inter-rater agreement. Radiology reporting should include clear definitions of clinically relevant nodal features as well as levels of certainty to serve various needs in clinical care and research.

Targeting fibroblast activation protein in newly diagnosed squamous cell carcinoma of the oral cavity - initial experience and comparison to [18F]FDG PET/CT and MRI

Purpose

While [¹⁸F]-fluorodeoxyglucose ([¹⁸F]FDG) is the standard for positron emission tomography/computed tomography (PET/CT) imaging of oral squamous cell carcinoma (OSCC), diagnostic specificity is hampered by uptake in inflammatory cells such as neutrophils or macrophages. Recently, molecular imaging probes targeting fibroblast activation protein α (FAP), which is overexpressed in a variety of cancer-associated fibroblasts, have become available and might constitute a feasible alternative to FDG PET/CT.

Methods

Ten consecutive, treatment-naïve patients (8 males, 2 females; mean age, 62 ± 9 years) with biopsy-proven OSCC underwent both whole-body [¹⁸F]FDG and [⁶⁸Ga]FAPI-04 (FAP-directed) PET/CT for primary staging prior to tumor resection and cervical lymph node dissection. Detection of the primary tumor, as well as the presence and number of lymph node and distant metastases was analysed. Intensity of tracer accumulation was assessed by means of maximum (SUV_max) and peak (SUV_peak) standardized uptake values. Histological work-up including immunohistochemical staining for FAP served as standard of reference.

Results

[¹⁸F]FDG and FAP-directed PET/CT detected all primary tumors with a SUV_max of 25.5 ± 13.2 (FDG) and 20.5 ± 6.4 (FAP-directed) and a SUV_peak of 16.1 ± 10.3 ([¹⁸F]FDG) and 13.8 ± 3.9 (FAP-directed), respectively. Regarding cervical lymph node metastases, FAP-directed PET/CT demonstrated comparable sensitivity (81.3% vs. 87.5%; P = 0.32) and specificity (93.3% vs. 81.3%; P = 0.16) to [¹⁸F]FDG PET/CT. FAP expression on the cell surface of cancer-associated fibroblasts in both primary lesions as well as lymph nodes metastases was confirmed in all samples.

Conclusion

FAP-directed PET/CT in OSCC seems feasible. Future research to investigate its potential to improve patient staging is highly warranted.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00259-021-05422-z.

Accuracy of 18-F Fluorodeoxyglucose Positron Emission Tomographic/Computed Tomographic Imaging in Primary Staging of Squamous Cell Carcinoma of the Oral Cavity

IMPORTANCE

Squamous cell carcinoma (SCC) of the oral cavity is one of the most common tumor entities worldwide. Precise initial staging is necessary to determine a diagnosis, treatment, and prognosis.

OBJECTIVE

To examine the diagnostic accuracy of preoperative 18-F fluorodeoxyglucose (FDG) positron emission tomographic/computed tomographic (PET/CT) imaging in detecting cervical lymph node metastases.

DESIGN, SETTING, AND PARTICIPANTS

This prospective diagnostic study was performed at a single tertiary reference center between June 1, 2013, and January 31, 2016. Data were analyzed from April 7, 2018, through May 31, 2019. Observers of the FDG PET/CT imaging were blinded to patients' tumor stage. A total of 150 treatment-naive patients with clinical suspicion of SCC of the oral cavity were enrolled.

EXPOSURES

All patients underwent FDG PET/CT imaging before local tumor resection with selective or complete neck dissection.

MAIN OUTCOMES AND MEASURES

The accuracy of FDG PET/CT in localizing primary tumor, lymph node, and distant metastases was tested. Histopathologic characteristics of the tissue samples served as the standard of reference.

RESULTS

Of the 150 patients enrolled, 135 patients (74 [54.8%] men) with a median age of 63 years (range, 23-88 years) met the inclusion criteria (histopathologically confirmed primary SCC of the oral cavity/level-based histopathologic assessment of the resected lymph nodes). Thirty-six patients (26.7%) in the study cohort had neck metastases. Use of FDG PET/CT detected cervical lymph node metastasis with 83.3% sensitivity (95% CI, 71.2%-95.5%) and 84.8% specificity (95% CI, 77.8%-91.9%) and had a negative predictive value of 93.3% (95% CI, 88.2%-98.5%). The specificity was higher than for contrast-enhanced cervical CT imaging (67.0%; 95% CI, 57.4%-76.7%; P < .01) and cervical magnetic resonance imaging (62.6%; 95% CI, 52.7%-72.6%; P < .001). Ipsilateral lymph node metastasis in left- or right-sided primary tumor sites was detected with 78.6% sensitivity (95% CI, 63.4%-93.8%) and 83.1% specificity (95% CI, 75.1%-91.2%), and contralateral metastatic involvement was detected with 66.7% sensitivity (95% CI, 28.9%-100.0%) and 98.6% specificity (95% CI, 95.9%-100.0%). No distant metastases were observed.

CONCLUSIONS AND RELEVANCE

In this study, FDG PET/CT imaging had a high negative predictive value in detecting cervical lymph node metastasis in patients with newly diagnosed, treatment-naive SCC of the oral cavity. Routine clinical use of FDG PET/CT might lead to a substantial reduction of treatment-related morbidity in most patients.

Actual applications of magnetic resonance imaging in dentomaxillofacial region

Magnetic resonance imaging (MRI) is a versatile imaging modality utilized in various medical fields. Specifically used for evaluation of soft tissues, with non-ionizing radiation and multiplanar sections that has provided great guidance to diagnosis. Nowadays, use of MRI in dental practice is becoming more pervasive, especially for the evaluation of head-and-neck cancer, detection of salivary gland lesions, lymphadenopathy, and temporomandibular joint disorders. Understanding the basic principles, its recent advances, and multiple applications in dentomaxillofacial region helps significantly in the diagnostic decision making. In this article, the principle of MRI and its recent advances are reviewed, with further discussion on the appearance of various maxillofacial pathosis.

Head and neck imaging surveillance strategy for HPV-positive oropharyngeal carcinoma following definitive (chemo)radiotherapy

PURPOSE

To describe the utilization pattern of head and neck (HN) surveillance imaging and explore the optimal strategy for radiologic "residual" lymph node (LN) surveillance following definitive (chemo)radiotherapy (RT/CRT) in human papillomavirus (HPV)+ oropharyngeal carcinoma (OPC).

METHODS

All HPV+ OPC patients who completed RT/CRT from 2012 to 2015 were included. Schedule and rationale for post-treatment HN-CT/MRI were recorded. Imaging findings and oncologic outcomes were evaluated.

RESULTS

A total of 1036 scans in 412 patients were reviewed: 414 scans for first post-treatment response assessment and 622 scans for the following reasons: follow-up of radiologic "residual" LN(s) (293 scans/175 patients); local symptoms (227/146); other (17/16); unknown (85/66). Rate of scans with "unstated" reason varied significantly among clinicians (3-28%, p < 0.001) and none of them yielded any positive imaging findings. First post-treatment scans identified 192 (47%) patients with radiologic "residual" LNs. Neck dissection (ND) was performed in 28 patients: 16 immediately (6/16 positive), 10 after one follow-up scan (2/10 positive), and 2 after 2nd follow-up scan (1/2 positive). Thirty patients had >2 consecutive follow-up scans at 2-3-month intervals, and none showed subsequent imaging progression or regional failure.

CONCLUSIONS

Pattern of HN imaging utilization for surveillance varied significantly among clinicians. Imaging surveillance reduces the need for ND. However, routine HN-CT/MR surveillance without clinical symptoms/signs does not demonstrate proven value in identifying locoregional failure or toxicity. Radiologic "residual" LNs without adverse features are common. If two subsequent follow-up scans demonstrate stable/regressing radiologic "residual" LNs, clinical surveillance without further imaging appears to be safe in this population.

Photoacoustic Molecular Imaging for the Identification of Lymph Node Metastasis in Head and Neck Cancer Using an Anti-EGFR Antibody-Dye Conjugate

The presence of lymph node (LN) metastases is an essential prognostic indicator in patients with head and neck squamous cell carcinoma (HNSCC). This study assessed photoacoustic molecular imaging (PAMI) of the antiepidermal growth factor receptor antibody (panitumumab) conjugated to a near-infrared fluorescent dye, IRDye800CW (panitumumab-IRDye800CW; pan800), for the identification of occult metastatic LNs in patients with HNSCC (n = 7). Methods: After in vitro photoacoustic imaging characterization of pan800, PAMI was performed on excised neck specimens from patients infused with pan800 before surgery. Freshly obtained neck specimens were imaged with 3-dimensional, multiwavelength spectroscopic PAMI (wavelengths of 680, 686, 740, 800, 860, 924, and 958 nm). Harvested LNs were then imaged with a closed-field near-infrared fluorescence imager and histologically examined by the pathologist to determine their metastatic status. Results: In total, 53 LNs with a maximum diameter of 10 mm were analyzed with photoacoustic and fluorescence imaging, of which 4 were determined to be metastatic on the final histopathologic report. Photoacoustic signals in the LNs corresponding to accumulated pan800 were spectrally unmixed using a linear least-square-error classification algorithm. The average thresholded photoacoustic signal intensity corresponding to pan800 was 5-fold higher for metastatic LNs than for benign LNs (2.50 ± 1.09 arbitrary units [a.u.] vs. 0.53 ± 0.32 a.u., P < 0.001). Fluorescence imaging showed that metastatic LNs had a 2-fold increase in fluorescence signal compared with benign LNs ex vivo (P < 0.01, 0.068 ± 0.027 a.u. vs. 0.035 ± 0.018 a.u.). Moreover, the ratio of the average of the highest 10% of the photoacoustic signal intensity over the total average, representative of the degree of heterogeneity in the pan800 signal in LNs, showed a significant difference between metastatic LNs and benign LNs (11.6 ± 13.4 vs. 1.8 ± 0.7, P < 0.01) and an area under the receiver-operating-characteristic curve of 0.96 (95% CI, 0.91-1.00). Conclusion: The data indicate that PAMI of IRDye800-labeled tumor-specific antibody may have the potential to identify occult LN metastasis perioperatively in HNSCC patients.

The use of structured reporting of head and neck ultrasound ensures time-efficiency and report quality during residency

PURPOSE

Free text reports (FTR) of head and neck ultrasound studies are currently deployed in most departments. Because of a lack of composition and language, these reports vary greatly in terms of quality and reliability. This may impair the learning process during residency. The purpose of the study was to analyze the longitudinal effects of using structured reports (SR) of head and neck ultrasound studies during residency.

METHODS

Attending residents (n = 24) of a tripartite course on head and neck ultrasound, accredited by the German Society for Ultrasound in Medicine (DEGUM), were randomly allocated to pictures of common diseases. Both SRs and FTRs were compiled. All reports were analyzed concerning completeness, acquired time and legibility. Overall user contentment was evaluated by a questionnaire.

RESULTS

SRs achieved significantly higher ratings regarding completeness (95.6% vs. 26.4%, p < 0.001), description of pathologies (72.2% vs. 58.9%, p < 0.001) and legibility (100% vs. 52.4%, p < 0.001) with a very high inter-rater reliability (Fleiss' kappa 0.9). Reports were finalized significantly faster (99.1 s vs. 115.0 s, p < 0.001) and user contentment was significantly better when using SRs (8.3 vs. 6.3, p < 0.001). In particular, only SRs showed a longitudinally increasing time efficiency (- 20.1 s, p = 0.036) while maintaining consistent completeness ratings.

CONCLUSIONS

The use of SRs of head and neck ultrasound studies results in an increased longitudinal time-efficiency while upholding the report quality at the same time. This may indicate an additive learning effect of structured reporting. Superior outcomes in terms of comprehensiveness, legibility and time-efficiency can be observed immediately after implementation.

Dual-Energy CT Texture Analysis With Machine Learning for the Evaluation and Characterization of Cervical Lymphadenopathy

Purpose

To determine whether machine learning assisted-texture analysis of multi-energy virtual monochromatic image (VMI) datasets from dual-energy CT (DECT) can be used to differentiate metastatic head and neck squamous cell carcinoma (HNSCC) lymph nodes from lymphoma, inflammatory, or normal lymph nodes.

Materials and methods

A retrospective evaluation of 412 cervical nodes from 5 different patient groups (50 patients in total) having undergone DECT of the neck between 2013 and 2015 was performed: (1) HNSCC with pathology proven metastatic adenopathy, (2) HNSCC with pathology proven benign nodes (controls for (1)), (3) lymphoma, (4) inflammatory, and (5) normal nodes (controls for (3) and (4)). Texture analysis was performed with TexRAD® software using two independent sets of contours to assess the impact of inter-rater variation. Two machine learning algorithms (Random Forests (RF) and Gradient Boosting Machine (GBM)) were used with independent training and testing sets and determination of accuracy, sensitivity, specificity, PPV, NPV, and AUC.

Results

In the independent testing (prediction) sets, the accuracy for distinguishing different groups of pathologic nodes or normal nodes ranged between 80 and 95%. The models generated using texture data extracted from the independent contour sets had substantial to almost perfect agreement. The accuracy, sensitivity, specificity, PPV, and NPV for correctly classifying a lymph node as malignant (i.e. metastatic HNSCC or lymphoma) versus benign were 92%, 91%, 93%, 95%, 87%, respectively.

Conclusion

Machine learning assisted-DECT texture analysis can help distinguish different nodal pathology and normal nodes with a high accuracy.

Radiomics and Artificial Intelligence for Biomarker and Prediction Model Development in Oncology

Unlabelled Image.

US-guided core needle biopsy under assistance of hydrodissection to diagnose small lymph node metastases adjacent to cervical large vessels

PURPOSE

We aimed to evaluate the safety and effectiveness of ultrasonography (US) guided core needle biopsy (CNB) with hydrodissection to diagnose small lymph node metastases adjacent to cervical large vessels.

METHODS

From January 2013 to October 2017, 31 patients with 31 cervical lymph node metastases adjacent to large vessels presented for US-guided CNB. The mean maximal diameter of lymph nodes was 0.93±0.16 cm (range, 0.6-1.2 cm). All patients underwent US-guided CNB with 18-gauge true-cut biopsy needle after hydrodissection with saline. The separation success rate (SSR) of the hydrodissection, technical success rate (TSR) of CNB, histopathologic success rate (HST), and complications were assessed.

RESULTS

The SSR of hydrodissection was 100% (31/31). After effective separation between the lymph node metastases and the adjacent large vessels with saline injection, the procedures of CNB were performed with a TSR of 100% (31/31). The HST of the lymph node metastases was 100% (31/31). Two patients complained of mild cervical swelling sensation during saline injection. No major complications such as injury of the large vessels or massive hemorrhage occurred.

CONCLUSION

Hydrodissection can facilitate safely and effectively US-guided CNB of subcentimeter cervical lymph nodes adjacent to large vessels, potentially impacting further therapeutic decisions.

Risk Factors Analysis of Pathologically Confirmed Cervical Lymph Nodes Metastasis in Oral Squamous Cell Carcinoma Patients with Clinically Negative Cervical Lymph Node: Results from a Cancer Center of Central China

Objective: To explore the risk factors of cervical lymph node metastasis in oral squamous cell carcinoma (OSCC) patients with clinical negative cervical lymph nodes(cN0) and provide a reference for clinical treatment.

Methods: The clinical data of 161 OSCC patients with cN0 were retrospectively analyzed. All patients underwent extended primary resection combined with cervical lymph node dissection. The level and number of cervical lymph node metastasis were confirmed by postoperative pathology. The risk factors of cervical lymph node metastasis in patients were analyzed by univariate and multivariate Logistic regression analysis.

Results: Thirty-one out of 161 cases (19%) were confirmed cervical lymph node metastasis. Among them, there were 28 cases of lymph node metastasis in one cervical level and 3 cases in two cervical levels. A total of 42 positive lymph nodes were detected in 34 cervical levels. The level number of positive areas in the IA, IB, IIA, IIB, III, IV and V levels was 2, 15, 12, 1, 4,0, and 0, respectively. The corresponding regional metastasis rates were 5.9%, 44.1%, 35.3%, 2.9%, 11.8%, 0% and 0%, respectively. The number of positive lymph node metastases in the corresponding levels were 2, 17, 17, 1, 5, 0, and 0 respectively. Univariate analysis showed that gender, age, lesion location, T stage, and perineural invasion/lymphvascular invasion (PNI/PVI) had no significant effect on cervical lymph node metastasis (P>0.05). The growth pattern, degree of differentiation, depth of invasion, neutrophil/lymphocyte ratio (NLR) and the short/long axis diameter ratio (S/L ratio) of lymph nodes were important factors influencing the cervical lymph node metastasis in cN0 OSCC patients (P<0.05). Multivariate Logistic regression analysis indicated that the growth pattern, degree of differentiation, depth of invasion, NLR, and the S/L ratio of lymph nodes were independent risk factors for cervical lymph node metastasis (P<0.05).

Conclusion: The growth pattern, degree of differentiation, depth of invasion, neutrophil/lymphocyte ratio, and the short/long axis diameter ratio of lymph nodes were the independent risk factors for pathological cervical lymph node metastasis in oral squamous cell carcinoma patients with cN0. If patients with the above risk factors receive nonstandard radical neck dissection or no dissection, it may be necessary for them to receive the corresponding regional postoperative radiotherapy.

Head and neck squamous cell carcinoma: prediction of cervical lymph node metastasis by dual-energy CT texture analysis with machine learning

OBJECTIVES

This study was conducted in order to evaluate a novel risk stratification model using dual-energy CT (DECT) texture analysis of head and neck squamous cell carcinoma (HNSCC) with machine learning to (1) predict associated cervical lymphadenopathy and (2) compare the accuracy of spectral versus single-energy (65 keV) texture evaluation for endpoint prediction.

METHODS

Eighty-seven patients with HNSCC were evaluated. Texture feature extraction was performed on virtual monochromatic images (VMIs) at 65 keV alone or different sets of multi-energy VMIs ranging from 40 to 140 keV, in addition to iodine material decomposition maps and other clinical information. Random forests (RF) models were constructed for outcome prediction with internal cross-validation in addition to the use of separate randomly selected training (70%) and testing (30%) sets. Accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were determined for predicting positive versus negative nodal status in the neck.

RESULTS

Depending on the model used and subset of patients evaluated, an accuracy, sensitivity, specificity, PPV, and NPV of up to 88, 100, 67, 83, and 100%, respectively, could be achieved using multi-energy texture analysis. Texture evaluation of VMIs at 65 keV alone or in combination with only iodine maps had a much lower accuracy.

CONCLUSIONS

Multi-energy DECT texture analysis of HNSCC is superior to texture analysis of 65 keV VMIs and iodine maps alone and can be used to predict cervical nodal metastases with relatively high accuracy, providing information not currently available by expert evaluation of the primary tumor alone.

KEY POINTS

• Texture features of HNSCC tumor are predictive of nodal status. • Multi-energy texture analysis is superior to analysis of datasets at a single energy. • Dual-energy CT texture analysis with machine learning can enhance noninvasive diagnostic tumor evaluation.

Impact of structured reporting on developing head and neck ultrasound skills

BACKGROUND

Reports of head and neck ultrasound examinations are frequently written by hand as free texts. This is a serious obstacle to the learning process of the modality due to a missing report structure and terminology. Therefore, there is a great inter-observer variability in overall report quality. Aim of the present study was to evaluate the impact of structured reporting on the learning process as indicated by the overall report quality of head and neck ultrasound examinations within medical school education.

METHODS

Following an immersion course on head and neck ultrasound, previously documented images of three common pathologies were handed out to 58 medical students who asked to create both standard free text reports (FTR) and structured reports (SR). A template for structured reporting of head and neck ultrasound examinations was created using a web-based approach. FTRs and SRs were evaluated with regard to overall quality, completeness, required time to completion and readability by two independent raters (Paired Wilcoxon test, 95% CI). Ratings were assessed for inter-rater reliability (Fleiss' kappa). Additionally, a questionnaire was utilized to evaluate user satisfaction.

RESULTS

SRs received significantly better ratings in terms of report completeness (97.7% vs. 53.5%, p < 0.001) regarding all items. In addition, pathologies were described in more detail using SRs (70% vs. 51.1%, p < 0.001). Readability was significantly higher in all SRs when compared to FTRs (100% vs. 54.4%, p < 0.001). Mean time to complete was significantly lower (79.6 vs. 205.4 s, p < 0.001) and user satisfaction was significantly higher when using SRs (8.5 vs. 4.1, p < 0.001). Also, inter-rater reliability was very high (Fleiss' kappa 0.93).

CONCLUSIONS

SRs of head and neck ultrasound examinations provide more detailed information with a better readability in a time-saving manner within medical education. Also, medical students may benefit from SRs in their learning process due to the structured approach and standardized terminology.

Structured reporting of head and neck ultrasound examinations

Background

Reports of head and neck ultrasound examinations are frequently written by hand as free texts. Naturally, quality and structure of free text reports is variable, depending on the examiner’s individual level of experience. Aim of the present study was to compare the quality of free text reports (FTR) and structured reports (SR) of head and neck ultrasound examinations.

Methods

Both standard FTRs and SRs of head and neck ultrasound examinations of 43 patients were acquired by nine independent examiners with comparable levels of experience. A template for structured reporting of head and neck ultrasound examinations was created using a web-based approach. FTRs and SRs were evaluated with regard to overall quality, completeness, required time to completion, and readability by four independent raters with different specializations (Paired Wilcoxon test, 95% CI) and inter-rater reliability was assessed (Fleiss’ kappa). A questionnaire was used to compare FTRs vs. SRs with respect to user satisfaction (Mann-Whitney U test, 95% CI).

Results

By comparison, completeness scores of SRs were significantly higher than FTRs’ completeness scores (94.4% vs. 45.6%, p < 0.001), and pathologies were described in more detail (91.1% vs. 54.5%, p < 0.001). Readability was significantly higher in all SRs when compared to FTRs (100% vs. 47.1%, p < 0.001). The mean time to complete a report, however, was significantly higher in SRs (176.5 vs. 107.3 s, p < 0.001). SRs achieved significantly higher user satisfaction ratings (VAS 8.87 vs. 1.41, p < 0.001) and a very high inter-rater reliability (Fleiss’ kappa 0.92).

Conclusions

As compared to FTRs, SRs of head and neck ultrasound examinations are more comprehensive and easier to understand. On the balance, the additional time needed for completing a SR is negligible. Also, SRs yield high inter-rater reliability and may be used for high-quality scientific data analyses.

Diagnosis of cervical lymph node metastases in head and neck cancer with ultrasonic measurement of lymph node volume

OBJECTIVE

The purpose of this study was to evaluate the usefulness of ultrasound (US) volume measurement of the cervical lymph nodes for diagnosing nodal metastasis in patients with head and neck cancer using a node-by-node comparison.

METHODS

Thirty-four consecutive patients with head and neck cancer from one tertiary university hospital were prospectively enrolled from 2012 to 2017. Patients with histologically proven squamous cell primary tumors in the head and neck region scheduled to undergo a therapeutic neck dissection were eligible. For each patient, 1-4 target lymph nodes were selected from the planned neck dissection levels. Lymph nodes with thickness >20 mm or in a cluster were excluded. Node-by-node comparisons between the pre-operative US assessment, the post-operative actual measurements and histopathological results were performed for all target lymph nodes. Quantitative measurements, such as three diameters, ratios of the three diameters and volume were analyzed in this study. Lymph node volume was calculated using the ellipsoid formula.

RESULTS

Patients comprised 28 men and 6 women with a mean age of 60.0 years (range, 29-80 years) at the time of surgery. In total, 67 target lymph nodes were analyzed in this study and the thickness ranged from 3.9 to 20.0 mm (mean 8.0 mm). There was a strong correlation between the US volume and post-operative actual volume (ρ = 0.87, p < 0.01). The US volume measured 2156 ± 2156 mm³ for the tumor positive nodes, which was significantly greater than the US volume of 512 ± 315 mm³ for tumor negative nodes (p < 0.01). Significant differences between tumor positive and tumor negative nodes were found in five variables (volume, thickness, major axis, minor axis and ratio of minor axis to thickness) for total lymph nodes. To identify predictors of lymph node metastasis, ROC curves of the US variables of target lymph nodes were compared, of which 4 variables were considered acceptable for predicting the lymph node metastasis: volume (AUC 0.86), thickness (AUC 0.86), major axis (AUC 0.79), and minor axis (AUC 0.79) for total lymph nodes. The optimal cut-off level for US volume in total lymph nodes was found to be 1242 mm³, whereby a 62% sensitivity and 98% specificity was reached (likelihood ratio: 25.2).

CONCLUSION

Pre-operative ultrasonic volume measurement of the cervical lymph nodes was useful for early detection of cervical nodal metastasis in head and neck cancer.

Dual-Energy CT Characteristics of Parathyroid Adenomas on 25-and 55-Second 4D-CT Acquisitions: Preliminary Experience

OBJECTIVE

The objective of this study was to compare the dual-energy computed tomography (CT) characteristics of parathyroid adenomas (PAs), thyroid tissue, and lymph nodes (LNs) and assess whether the spectral information can improve distinction of these tissues.

METHODS

Dual-energy CT scans from 20 patients with pathologically proven PAs were retrospectively evaluated, identifying 19 eligible PAs and region of interest analysis used for spectral characterization.

RESULTS

There was a significant difference in multiple spectral parameters between PAs, LNs, and the thyroid gland (P < 0.05-0.0001). The greatest difference in spectral characteristics of PAs compared with that of LNs was on the 25-second acquisition, whereas the 55-second acquisition was better for distinguishing PAs from the thyroid gland.

CONCLUSIONS

Four-dimensional CT acquired in dual-energy CT mode has the potential to further enhance diagnostic accuracy for PA identification on individual phases of the perfusion study.

Differentiation of malignant cervical lymphadenopathy by dual-energy CT: a preliminary analysis

The accurate diagnosis of malignant cervical lymphadenopathy remains challenging. In this study, we determined the value of quantitative parameters derived from dual-energy computed tomography (DECT) for differentiating malignant cervical lymphadenopathy caused by thyroid carcinoma (TC), salivary gland carcinoma (SC), squamous cell carcinoma (SCC) and lymphoma. We retrospectively analysed 92 patients with pathologically confirmed cervical lymphadenopathy due to TC, SC, SCC and lymphoma. All patients received a DECT scan before therapy. Using GSI (gemstone spectral imaging) Volume Viewer software, we analysed the enhanced monochromatic data, and the quantitative parameters we acquired included the iodine concentration (IC), water concentration (WC) and the slope of the spectral HU curve (λ_HU). One-way ANOVA showed significant differences in the IC and λ_HU among different groups (P < 0.05). Post-hoc pairwise comparisons demonstrated the IC and λ_HU of TC group were significantly higher than those of SC, SCC and lymphoma groups (P < 0.05). In addition, the IC and λ_HU of SC group were significantly higher than those of the SCC and lymphoma groups (P < 0.05). Other comparisons of IC and λ_HU values showed no significant differences (P > 0.05). The quantitative parameters derived from DECT were useful supplements to conventional computed tomography images and were helpful for distinguishing different malignant cervical lymphadenopathies.

Optimal Virtual Monochromatic Images for Evaluation of Normal Tissues and Head and Neck Cancer Using Dual-Energy CT

BACKGROUND AND PURPOSE

Dual-energy CT is not used routinely for evaluation of the head and neck, and there is no consensus on the optimal virtual monochromatic image energies for evaluating normal tissues or head and neck cancer. We performed a quantitative evaluation to determine the optimal virtual monochromatic images for visualization of normal tissues, head and neck squamous cell carcinoma, and lymphadenopathy.

MATERIALS AND METHODS

Dual-energy CT scans from 10 healthy patients and 30 patients with squamous cell carcinoma were evaluated at different virtual monochromatic energy levels ranging from 40 to 140 keV. The signal-to-noise ratios of muscles at 6 different levels, glands (parotid, sublingual, submandibular, and thyroid), 30 tumors, and 17 metastatic lymph nodes were determined as measures of optimal image quality. Lesion attenuation and contrast-to-noise ratios (compared with those of muscle) were evaluated to assess lesion conspicuity.

RESULTS

The optimal signal-to-noise ratio for all the tissues was at 65 keV (P < .0001). However, tumor attenuation (P < .0001), attenuation difference between tumor and muscles (P = .03), and lesion contrast-to-noise ratios (P < .0001) were highest at 40 keV.

CONCLUSIONS

The optimal image signal-to-noise ratio is at 65 keV, but tumor conspicuity compared with that of muscle is greatest at 40 keV. Optimal evaluation of the neck may be best achieved by a multiparametric approach, with 65-keV virtual monochromatic images providing the best overall image quality and targeted use of 40-keV virtual monochromatic images for tumor evaluation.