An MRI-based Scoring System for Preoperative Prediction of Axillary Response to Neoadjuvant Chemotherapy in Node-Positive Breast Cancer: A Multicenter Retrospective Study.

RATIONALE AND OBJECTIVES: To develop and validate a simplified scoring system by integrating MRI and clinicopathologic features for preoperative prediction of axillary pathologic complete response (pCR) to neoadjuvant chemotherapy (NAC) in clinically node-positive breast cancer. MATERIALS AND METHODS: A total of 389 patients from three hospitals were retrospectively analyzed. To identify independent predictors for axillary pCR, univariable and multivariable logistic regression analyses were performed on pre- and post-NAC MRI and clinicopathologic features. Then, a simplified scoring system was constructed based on regression coefficients of predictors in the multivariable model, and its predictive performance was assessed with the receiver operating characteristic curve and calibration curve. The added value of the scoring system for reducing false-negative rate (FNR) of the sentinel lymph node biopsy (SLNB) was also evaluated. RESULTS: The simplified scoring system including seven predictors: progesterone receptor-negative (Three points), HER2-positive (Two points), post-NAC clinical T0-1 stage (Two points), pre-NAC higher ADC value of breast tumor (One point), absence of perinodal infiltration at pre-NAC (One point) and post-NAC MRI (Two points), and absence of enhancement in the tumor bed at post-NAC MRI (Two points), showed good calibration and discrimination, with AUCs of 0.835, 0.828 and 0.798 in the training, internal and external validation cohorts, respectively. The axillary pCR rates were increased with the total points of the scoring system, and patients with a score of ≥11 points had a pCR rate of 86%-100%. In test cohorts for simulating clinical application, the diagnostic accuracy for axillary pCR was 80%-90% among four different radiologists. Compared to standalone SLNB, combining the scoring system with SLNB reduced the FNR from 14.5% to 4.8%. CONCLUSION: The clinicopathologic-image scoring system with good predictive performance for axillary pCR in clinically node-positive breast cancer, may guide axillary management after NAC and improve patient selection for de-escalating axillary surgery to reduce morbidity.

Multi-layer pseudo-supervision for histopathology tissue semantic segmentation using patch-level classification labels.

Tissue-level semantic segmentation is a vital step in computational pathology. Fully-supervised models have already achieved outstanding performance with dense pixel-level annotations. However, drawing such labels on the giga-pixel whole slide images is extremely expensive and time-consuming. In this paper, we use only patch-level classification labels to achieve tissue semantic segmentation on histopathology images, finally reducing the annotation efforts. We propose a two-step model including a classification and a segmentation phases. In the classification phase, we propose a CAM-based model to generate pseudo masks by patch-level labels. In the segmentation phase, we achieve tissue semantic segmentation by our propose Multi-Layer Pseudo-Supervision. Several technical novelties have been proposed to reduce the information gap between pixel-level and patch-level annotations. As a part of this paper, we introduce a new weakly-supervised semantic segmentation (WSSS) dataset for lung adenocarcinoma (LUAD-HistoSeg). We conduct several experiments to evaluate our proposed model on two datasets. Our proposed model outperforms five state-of-the-art WSSS approaches. Note that we can achieve comparable quantitative and qualitative results with the fully-supervised model, with only around a 2% gap for MIoU and FwIoU. By comparing with manual labeling on a randomly sampled 100 patches dataset, patch-level labeling can greatly reduce the annotation time from hours to minutes. The source code and the released datasets are available at: https://github.com/ChuHan89/WSSS-Tissue.

A Nomogram Based on CT Deep Learning Signature: A Potential Tool for the Prediction of Overall Survival in Resected Non-Small Cell Lung Cancer Patients.

PURPOSE: To develop and further validate a deep learning signature-based nomogram from computed tomography (CT) images for prediction of the overall survival (OS) in resected non-small cell lung cancer (NSCLC) patients. PATIENTS AND METHODS: A total of 1792 deep learning features were extracted from non-enhanced and venous-phase CT images for each NSCLC patient in training cohort (n=231). Then, a deep learning signature was built with the least absolute shrinkage and selection operator (LASSO) Cox regression model for OS estimation. At last, a nomogram was constructed with the signature and other independent clinical risk factors. The performance of nomogram was assessed by discrimination, calibration and clinical usefulness. In addition, in order to quantify the improvement in performance added by deep learning signature, the net reclassification improvement (NRI) was calculated. The results were validated in external validation cohort (n=77). RESULTS: A deep learning signature with 9 selected features was significantly associated with OS in both training cohort (hazard ratio [HR]=5.455, 95% CI: 3.393-8.769, P<0.001) and external validation cohort (HR=3.029, 95% CI: 1.673-5.485, P=0.004). The nomogram combining deep learning signature with clinical risk factors of TNM stage, lymphatic vessel invasion and differentiation grade showed favorable discriminative ability with C-index of 0.800 as well as a good calibration, which was validated in external validation cohort (C-index=0.723). Additional value of deep learning signature to the nomogram was statistically significant (NRI=0.093, P=0.027 for training cohort; NRI=0.106, P=0.040 for validation cohort). Decision curve analysis confirmed the clinical usefulness of this nomogram in predicting OS. CONCLUSION: The deep learning signature-based nomogram is a robust tool for prognostic prediction in resected NSCLC patients.

Radiomic Nomogram for Pretreatment Prediction of Pathologic Complete Response to Neoadjuvant Therapy in Breast Cancer: Predictive Value of Staging Contrast-enhanced CT.

INTRODUCTION: The purpose of this study was to predict pathologic complete response (pCR) to neoadjuvant therapy in breast cancer using radiomics based on pretreatment staging contrast-enhanced computed tomography (CECT). PATIENTS AND METHODS: A total of 215 patients were retrospectively analyzed. Based on the intratumoral and peritumoral regions of CECT images, radiomic features were extracted and selected, respectively, to develop an intratumoral signature and a peritumoral signature with logistic regression in a training dataset (138 patients from November 2015 to October 2017). We also developed a clinical model with the molecular characterization of the tumor. A radiomic nomogram was further constructed by incorporating the intratumoral and peritumoral signatures with molecular characterization. The performance of the nomogram was validated in terms of discrimination, calibration, and clinical utility in an independent validation dataset (77 patients from November 2017 to December 2018). Stratified analysis was performed to develop a subtype-specific radiomic signature for each subgroup. RESULTS: Compared with the clinical model (area under the curve [AUC], 0.756), the radiomic nomogram (AUC, 0.818) achieved better performance for pCR prediction in the validation dataset with continuous net reclassification improvement of 0.787 and good calibration. Decision curve analysis suggested the nomogram was clinically useful. Subtype-specific radiomic signatures showed improved AUCs (luminal subgroup, 0.936; human epidermal growth factor receptor 2-positive subgroup, 0.825; and triple negative subgroup, 0.858) for pCR prediction. CONCLUSION: This study has revealed a predictive value of pretreatment staging-CECT and successfully developed and validated a radiomic nomogram for individualized prediction of pCR to neoadjuvant therapy in breast cancer, which could assist clinical decision-making and improve patient outcome.