Dataset Information

Deep learning identifies morphological patterns of homologous recombination deficiency in luminal breast cancers from whole slide images.

ABSTRACT: Homologous recombination DNA-repair deficiency (HRD) is becoming a well-recognized marker of platinum salt and polyADP-ribose polymerase inhibitor chemotherapies in ovarian and breast cancers. While large-scale screening for HRD using genomic markers is logistically and economically challenging, stained tissue slides are routinely acquired in clinical practice. With the objectives of providing a robust deep-learning method for HRD prediction from tissue slides and identifying related morphological phenotypes, we first show that digital pathology workflows are sensitive to potential biases in the training set, then we propose a method to overcome the influence of these biases, and we develop an interpretation method capable of identifying complex phenotypes. Application to our carefully curated in-house dataset allows us to predict HRD with high accuracy (area under the receiver-operator characteristics curve 0.86) and to identify morphological phenotypes related to HRD. In particular, the presence of laminated fibrosis and clear tumor cells associated with HRD open new hypotheses regarding its phenotypic impact.

SUBMITTER: Lazard T

PROVIDER: S-EPMC9798078 | biostudies-literature | 2022 Dec

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Deep learning identifies morphological patterns of homologous recombination deficiency in luminal breast cancers from whole slide images.

Lazard Tristan T Bataillon Guillaume G Naylor Peter P Popova Tatiana T Bidard François-Clément FC Stoppa-Lyonnet Dominique D Stern Marc-Henri MH Decencière Etienne E Walter Thomas T Vincent-Salomon Anne A

Cell reports. Medicine 20221213 12

Homologous recombination DNA-repair deficiency (HRD) is becoming a well-recognized marker of platinum salt and polyADP-ribose polymerase inhibitor chemotherapies in ovarian and breast cancers. While large-scale screening for HRD using genomic markers is logistically and economically challenging, stained tissue slides are routinely acquired in clinical practice. With the objectives of providing a robust deep-learning method for HRD prediction from tissue slides and identifying related morphologic ...[more]

PMID: 36516847

Similar Datasets

Project description:BackgroundThe homologous recombination deficiency (HRD) test is an important tool for identifying patients with epithelial ovarian cancer (EOC) benefit from the treatment with poly(adenosine diphosphate-ribose) polymerase inhibitor (PARPi). Using whole exome sequencing (WES)-based platform can provide information of gene mutations and HRD score; however, the clinical value of WES-based HRD test was less validated in EOC.MethodsWe enrolled 40 patients with EOC in the training cohort and 23 in the validation cohort. The WES-based HRD score was calculated using the scarHRD software. We first evaluated the concordance of the HRD status defined by the Myriad MyChoice CDx and then assessed the value of HRD on clinical prognosis in patients with EOC.ResultsThe HRD score defined by the WES-based test was positively correlated with that of the Myriad MyChoice® CDx test (r = 0.82, p < 0.01) in the training cohort. In compared to HRD status of Myriad test, the sensitivity, specificity, positive predictive value, and negative predictive value of the WES-based HRD test were 93.5% (29/31), 77.8% (7/9), 93.5% (29/31), and 77.8% (7/9), respectively. Patients with positive HRD status defined by WES-based scarHRD test and Myriad MyChoice® CDx test were both highly associated with platinum sensitive response (both Fisher's exact test, p = 0.002) as well as the superior progression-free survival (both log-rank p = 0.002). The multi-variate Cox regression model incorporated with optimal debulking surgery showed that the recurrence risk was decreased in the patients with positive HRD status, either defined by Myriad MyChoice® CDx test (Hazard ratio (HR) 0.33, 95% confidence interval (CI) 0.14-0.79, p = 0.013) or WES-based test Myriad MyChoice® CDx test (HR 0.34, 95% CI 0.14-0.80, p = 0.014). Nine patients had mutations in the genes involved in HR DNA repair, and all of them were positive for HRD. In the validation group, 23 patients were defined as positive HRD by WES-based testing. Six positive HRD patients and 5 negative HRD patients received maintenance PARPi. The median responsive interval of PARPi was 17 months in positive HRD patients and 3 months in negative HRD patients.ConclusionThe WES-based test is a potential option for determining the HRD status in EOC patients, and desires for further validation in large-scale cohorts.

Dataset Information

Deep learning identifies morphological patterns of homologous recombination deficiency in luminal breast cancers from whole slide images.

Publications

Deep learning identifies morphological patterns of homologous recombination deficiency in luminal breast cancers from whole slide images.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets