Search

After chemotherapy, patients with non-seminomatous germ cell tumors (NSGCTs) with residual masses >1 cm on computed tomography (CT) undergo surgery. However, in approximately 50% of cases, these masses only consist of necrosis/fibrosis. We aimed to develop a radiomics score to predict the malignant character of residual masses to avoid surgical overtreatment. Patients with NSGCTs who underwent surgery for residual masses between September 2007 and July 2020 were retrospectively identified from a unicenter database. Residual masses were delineated on post-chemotherapy contrast-enhanced CT scans. Tumor textures were obtained using the free software LifeX. We constructed a radiomics score using a penalized logistic regression model in a training dataset, and evaluated its performance on a test dataset. We included 76 patients, with 149 residual masses; 97 masses were malignant (65%). In the training dataset (n = 99 residual masses), the best model (ELASTIC-NET) led to a radiomics score based on eight texture features. In the test dataset, the area under the curve (AUC), sensibility, and specificity of this model were respectively estimated at 0.82 (95%CI, 0.69–0.95), 90.6% (75.0–98.0), and 61.1% (35.7–82.7). Our radiomics score may help in the prediction of the malignant nature of residual post-chemotherapy masses in NSGCTs before surgery, and thus limit overtreatment. However, these results are insufficient to simply select patients for surgery.

Purpose The purpose of this study was to create a deep learning algorithm to infer the benign or malignant nature of breast nodules using two-dimensional B-mode ultrasound data initially marked as BI-RADS 3 and 4. Materials and methods An ensemble of mask region-based convolutional neural networks (Mask-RCNN) combining nodule segmentation and classification were trained to explicitly localize the nodule and generate a probability of the nodule to be malignant on two-dimensional B-mode ultrasound. These probabilities were aggregated at test time to produce final results. Resulting inferences were assessed using area under the curve (AUC). Results A total of 460 ultrasound images of breast nodules classified as BI-RADS 3 or 4 were included. There were 295 benign and 165 malignant breast nodules used for training and validation, and another 137 breast nodules images used for testing. As a part of the challenge, the distribution of benign and malignant breast nodules in the test database remained unknown. The obtained AUC was 0.69 (95% CI: 0.57–0.82) on the training set and 0.67 on the test set. Conclusion The proposed deep learning solution helps classify benign and malignant breast nodules based solely on two-dimensional ultrasound images initially marked as BIRADS 3 and 4.

This paper deals with the place in 2021 of artificial intelligence for screening of breast cancer in France et discusses of the different types of use and their performance in the literature (Standalone, Augmented radiologist, triage). Moreover, this paper will give an overview of the potential applications of AI in the future (Personalized screening, prediction of interval cancers).

Second reading is an important part of breast cancer organized screening program. Image quality control and detection of non-diagnosed cancer by first reader are the two goals of this process. In France, 6 % of all screening cancer are diagnosed by second reading, actually done on screen film. With the technologic evolution (Digital breast tomosynthesis, Artificial intelligence) and societal digitalization, this process need to evolve. After some report about organization and results for second reading in France and outside, current and future shortcomings, proposition from professionals involved in breast cancer screening are made to improve this public health program.

Purpose :Since ACOSOG Z1011 trial, the role of axillary US preoperative has been redirected for predicting extensive axillary involvement. Our study is aimed to evaluate routinely US axillary prediction for extensive nodal burden in breast cancer. Materials and methods :The study was approved by the institutional reaview board and the requirement for in-formed consent was waived.All patients in our institution with early-stage invasive breast cancer have a systematic axillary ultrasound (US) evaluation before surgery.A prospective study of radiology and pathology records concerns all consecutives patients to evaluate US correlation with surgical findings beetwen November 1, 2020 and July 15, 2021.Neoadjuvant chemotherapy, ductal carcinoma in situ on breast biopsy, and relapse cases were excluded.Criteria used to define abnormal lymph nodes included : focal or diffuse cortical thickening with cortex thicker than >4mm, replacement or effacement of the fatty hilium, or rounder lymph node.3 US prediction groups were defined as: US G0 normal (no suspicious lymph node) ;US G1 non extensive involvement (1 or 2 suspicious nodes) ; US G2 extensive involvement (≥3 suspicious nodes).In case of preoperative positive axillary biopsy, positive intraoperative examination, or patients without Acosog criteria, an axillary dissection was systematically performed.3 pathologic groups were defined as : path G0 (pN0) ; path G1 (≤2pN+) ; path G2 (≥3pN+).Statiscal analysis was performed with the chi-square the Fisher’s exact test.NPV for extensive axillary burden is defined as the proportion of patients with a negative test result (US G0 or US G1) that does not have extensive nodal involvement histologically (path G0 or path G1). Results. The study is still ongoing. On the 241 first cases, 64.3% (155/241) were cT1, 82.9% (200/241) were cN0. 163(78.3%) had Ductal invasive carcinoma, 37 were lobular histologic characteristics (17.8%) and 8 were mixted (3.8%).6.2% (15/241) patients presented extensive nodal involvement. ConclusionThe results of this study are preliminary. They will be completed after the next study closure. In the first results, it is globally observed that the preoperative ultrasound makes it possible to exclude more than 97% of extensive involvement (>2pN+). In case of less than 3 US suspected axillary nodes , sentinel node biospy remains indicated. The false negative rate cannot be calculated on this day and will not be presented until later. Path G0Path G1Path G2totalUS G0156243187US G1310922US G20033total1593415241VPN for extensive nodal disease is 97.9% (193/197).In abnormal US groups (US G1+US G2), 48% (12/24) patients had extensive axillary involvement. Citation Format: Camille Defoort, Joanna Chemaly, Luc Ceugnart, Marie-Pierre Chauvet. Axillary extensive burden evaluation by ultrasound in preoperative nodal staging in breast cancer patients [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr P1-01-11.