Dataset Information

Hybridizing mechanistic modeling and deep learning for personalized survival prediction after immune checkpoint inhibitor immunotherapy.

ABSTRACT: We present a study where predictive mechanistic modeling is combined with deep learning methods to predict individual patient survival probabilities under immune checkpoint inhibitor (ICI) immunotherapy. This hybrid approach enables prediction based on both measures that are calculable from mechanistic models of key mechanisms underlying ICI therapy that may not be directly measurable in the clinic and easily measurable quantities or patient characteristics that are not always readily incorporated into predictive mechanistic models. A deep learning time-to-event predictive model trained on a hybrid mechanistic + clinical data set from 93 patients achieved higher per-patient predictive accuracy based on event-time concordance, Brier score, and negative binomial log-likelihood-based criteria than when trained on only mechanistic model-derived values or only clinical data. Feature importance analysis revealed that both clinical and model-derived parameters play prominent roles in increasing prediction accuracy, further supporting the advantage of our hybrid approach.

SUBMITTER: Butner JD

PROVIDER: S-EPMC11324794 | biostudies-literature | 2024 Aug

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Hybridizing mechanistic modeling and deep learning for personalized survival prediction after immune checkpoint inhibitor immunotherapy.

Butner Joseph D JD Dogra Prashant P Chung Caroline C Koay Eugene J EJ Welsh James W JW Hong David S DS Cristini Vittorio V Wang Zhihui Z

NPJ systems biology and applications 20240814 1

We present a study where predictive mechanistic modeling is combined with deep learning methods to predict individual patient survival probabilities under immune checkpoint inhibitor (ICI) immunotherapy. This hybrid approach enables prediction based on both measures that are calculable from mechanistic models of key mechanisms underlying ICI therapy that may not be directly measurable in the clinic and easily measurable quantities or patient characteristics that are not always readily incorporat ...[more]

PMID: 39143136

Similar Datasets

Project description:BackgroundLung cancer (LC) is a leading cause of cancer-related mortality, and immunotherapy (IO) has shown promise in treating advanced-stage LC. However, identifying patients likely to benefit from IO and monitoring treatment response remains challenging. This study aims to develop a predictive model for progression-free survival (PFS) in LC patients with IO based on clinical features and advanced imaging biomarkers.Materials and methodsA retrospective analysis was conducted on a cohort of 206 LC patients receiving IO treatment. Pre-treatment computed tomography images were used to extract advanced imaging biomarkers, including intratumoral and peritumoral-vasculature radiomics. Clinical features, including age, gene status, hematology, and staging, were also collected. Key radiomic and clinical features for predicting IO outcomes were identified using a two-step feature selection process, including univariate Cox regression and chi-squared test, followed by sequential forward selection. The DeepSurv model was constructed to predict PFS based on clinical and radiomic features. Model performance was evaluated using the area under the time-dependent receiver operating characteristic curve (AUC) and concordance index (C-index).ResultsCombining radiomics of intratumoral heterogeneity and peritumoral-vasculature with clinical features demonstrated a significant enhancement (p < 0.001) in predicting IO response. The proposed DeepSurv model exhibited a prediction performance with AUCs ranging from 0.76 to 0.80 and a C-index of 0.83. Furthermore, the predicted personalized PFS curves revealed a significant difference (p < 0.05) between patients with favorable and unfavorable prognoses.ConclusionsIntegrating intratumoral and peritumoral-vasculature radiomics with clinical features enabled the development of a predictive model for PFS in LC patients with IO. The proposed model's capability to estimate individualized PFS probability and differentiate the prognosis status held promise to facilitate personalized medicine and improve patient outcomes in LC.

Project description:BackgroundNeoadjuvant checkpoint inhibition (CPI) has recently demonstrated impressive outcomes in patients with stage 3 cutaneous melanoma. However, the safety, efficacy, and outcome of neoadjuvant CPI in patients with mucosal melanoma (MM) are not well studied as MM is a rare melanoma subtype. CPI such as combination nivolumab and ipilimumab achieves response rates of 37-43% in unresectable or metastatic MM but there is limited data regarding the efficacy of these agents in the preoperative setting. We hypothesize that neoadjuvant CPI is a safe and feasible approach for patients with resectable MM.MethodUnder an institutionally approved protocol, we identified adult MM patients with resectable disease who received neoadjuvant anti-PD1 +/- anti-CTLA4 between 2015 to 2019 at our institution. Clinical information include age, gender, presence of nodal involvement or satellitosis, functional status, pre-treatment LDH, tumor mutation status, and treatment data was collected. Outcomes include event free survival (EFS), overall survival (OS), objective response rate (ORR), pathologic response rate (PRR), and grade ≥3 toxicities.ResultsWe identified 36 patients. Median age was 62; 58% were female. Seventy-eight percent of patients received anti-PD1 + anti-CTLA4. Node positive disease or satellite lesions was present at the time of treatment initiation in 47% of patients. Primary sites of disease were anorectal (53%), urogenital (25%), head and neck (17%), and esophageal (6%). A minority of patients did not undergo surgery due to complete response (n=3, 8%) and disease progression (n=6, 17%), respectively. With a median follow up of 37.9 months, the median EFS was 9.2 months with 3-year EFS rate of 29%. Median OS had not been reached and 3-year OS rate was 55%. ORR was 47% and PRR was 35%. EFS was significantly higher for patients with objective response and for patients with pathologic response. OS was significantly higher for patients with pathologic response. Grade 3 toxicities were reported in 39% of patients.ConclusionNeoadjuvant CPI for resectable MM is a feasible approach with signs of efficacy and an acceptable safety profile. As there is currently no standard approach for resectable MM, this study supports further investigations using neoadjuvant therapy for these patients.

Dataset Information

Hybridizing mechanistic modeling and deep learning for personalized survival prediction after immune checkpoint inhibitor immunotherapy.

Publications

Hybridizing mechanistic modeling and deep learning for personalized survival prediction after immune checkpoint inhibitor immunotherapy.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets