Survival Prediction In Lung Cancer Research Articles

Abstract 4905: Multimodal transformer model improves survival prediction in lung cancer compared to unimodal approaches

Abstract Integrating multimodal lung data including clinical notes, medical images, and molecular data is critical for predictive modeling tasks like survival prediction, yet effectively aligning these disparate data types remains challenging. We present a novel method to integrate heterogeneous lung modalities by first thoroughly analyzing various domain-specific models and selecting the optimal model for embedding feature extraction per data type based on performance on representative pretrained tasks. For clinical notes, the GatorTron models showed the lowest regression loss on an initial evaluation set, with the large GatorTron-medium model achieving 12.9 loss. After selecting the top performers, we extracted robust embeddings on the full lung dataset built using the Multimodal Integration of Oncology Data System (MINDS) framework. MINDS provides an end-to-end platform for aggregating and normalizing multimodal patient data. We aligned the multimodal embeddings to a central pre-trained language model using contrastive representation learning based on a cosine similarity loss function. To adapt the language model to the new modalities, we employed a parameter-efficient tuning method called adapter tuning, which introduces small trainable adapter layers that leave the base model weights frozen. This avoids catastrophic forgetting of the pretrained weights. We evaluated our multimodal model on prognostic prediction tasks including survival regression and subtype classification using both public and internal lung cancer datasets spanning multiple histologic subtypes and stages. Our aligned multimodal model demonstrated improved performance over models utilizing only single modalities, highlighting the benefits of integrating complementary information across diverse lung data types. This work illustrates the potential of flexible multimodal modeling for critical lung cancer prediction problems using heterogeneous real-world patient data. Our model provides a strong foundation for incorporating emerging data types, modalities, and predictive tasks in the future. Citation Format: Aakash Tripathi, Asim Waqas, Yasin Yilmaz, Ghulam Rasool. Multimodal transformer model improves survival prediction in lung cancer compared to unimodal approaches [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 4905.

Predicting Lung Cancer Survivability: A Machine Learning Ensemble Method On Seer Data

Ensemble methods are powerful techniques used in machine learning to improve the prediction accuracy of classifier learning systems. In this study, different ensemble learning methods for lung cancer survival prediction were evaluated on the Surveillance, Epidemiology and End Results (SEER) dataset. Data were preprocessed in several steps before applying classification models. The popular ensemble methods Bagging, Adaboost and three classification algorithms, K-Nearest Neighbours, Decision Tree and Neural Networks as base classifiers were evaluated for lung cancer survival prediction. The results empirically showed that ensemble methods are able to evaluate the performance of their base classifiers and they are appropriate methods for analysis of cancer survival.

Lung cancer survival prediction and biomarker identification with an ensemble machine learning analysis of tumor core biopsy metabolomic data.

While prediction of short versus long term survival from lung cancer is clinically relevant in the context of patient management and therapy selection, it has proven difficult to identify reliable biomarkers of survival. Metabolomic markers from tumor core biopsies have been shown to reflect cancer metabolic dysregulation and hold prognostic value. Implement and validate a novel ensemble machine learning approach to evaluate survival based on metabolomic biomarkers from tumor core biopsies. Data were obtained from tumor core biopsies evaluated with high-resolution 2DLC-MS/MS. Unlike biofluid samples, analysis of tumor tissue is expected to accurately reflect the cancer metabolism and its impact on patient survival. A comprehensive suite of machine learning algorithms were trained as base learners and then combined into a stacked-ensemble meta-learner for predicting "short" versus "long" survival on an external validation cohort. An ensemble method of feature selection was employed to find a reliable set of biomarkers with potential clinical utility. Overall survival (OS) is predicted in external validation cohort with AUROCTEST of 0.881 with support vector machine meta learner model, while progression-free survival (PFS) is predicted with AUROCTEST of 0.833 with boosted logistic regression meta learner model, outperforming a nomogram using covariate data (staging, age, sex, treatment vs. non-treatment) as predictors. Increased relative abundance of guanine, choline, and creatine corresponded with shorter OS, while increased leucine and tryptophan corresponded with shorter PFS. In patients that expired, N6,N6,N6-Trimethyl-L-lysine, L-pyrogluatmic acid, and benzoic acid were increased while cystine, methionine sulfoxide and histamine were decreased. In patients with progression, itaconic acid, pyruvate, and malonic acid were increased. This study demonstrates the feasibility of an ensemble machine learning approach to accurately predict patient survival from tumor core biopsy metabolomic data.

Automatic Feature Construction Based on Genetic Programming for Survival Prediction in Lung Cancer Using CT Images.

In the radiomics workflow, machine learning builds classification models from a set of input features. However, some features can be irrelevant and redundant, reducing the classification performance. This paper proposes using the Genetic Programming (GP) algorithm to automatically construct a reduced number of independent and relevant radiomic features. The proposed method is applied to patients affected by Non-Small Cell Lung Cancer (NSCLC) with pre-operative computed tomography (CT) images to predict the two-year survival by the use of linear classifiers. The model built using GP features is compared with benchmark models built using traditional features. The use of the GP algorithm increased classification performance: [Formula: see text] for the proposed model vs. [Formula: see text] and 0.64 for the benchmark models. Hence, the proposed approach better stratifies patients at high and low risk according to their overall postoperative survival time.

PO-1763 Radiomics and Deep Learning for the 2-Year Overall Survival Prediction in Lung Cancer

PO-1763 Radiomics and Deep Learning for the 2-Year Overall Survival Prediction in Lung Cancer

MA13.01 A Validation Study on DNA Repair Gene Variant for Lung Cancer Survival Prediction after Chemoradiation: A Secondary Analysis for RTOG-0617 Study

We previously reported that important genetic variants in DNA repair pathway genes particularly rs238406 in ERCC2 may be a predictive biomarker of radiation dose in patients with locally-advanced non-small cell lung cancer (LA-NSCLC) treated with chemoradiation (Wang et al, 2012; Jin et al, 2015). This study aimed to validate these findings using an independent cohort of NSCLC patients treated with high and standard doses of radiotherapy (RT). The study population was from RTOG-0617 clinical trial, a phase III trial comparing overall survival between high and standard RT doses with concurrent chemotherapy, with or without cetuximab, in stage III NSCLC. Patients with blood samples available for analysis were eligible to this study. DNA samples for genotyping were extracted from buffy-coat that were collected before commencement of treatment. Over two dozens functionally important SNPs of DNA repair pathway were tested using MassArray System. According to our previous study, this validation study focused solely on rs11615, rs3212948 in ERCC1 and rs238406 in ERCC2 in the nucleotide excision repair (NER) pathway, which were significant prognostic factors with respect to survival. The primary endpoint was OS, calculated from randomization to death due to any cause or last follow-up. Log-rank test and Cox proportional hazards model were used to examine the effects of genotypes on OS. The interaction term between respective SNP and RT dose groups was of primary interest. 321 patients met the selection criteria. Similar to findings of our previous study, Cox models found neither rs11615 nor rs3212948 in ERCC1 had a significant interaction effect with RT dose group (P>0.05). Notably, there was a statistically significant interaction between ERCC2: rs238406 (GT/TT vs. GG) and radiation dose (P for interaction=0.011). Specifically, in standard dose RT group patients carrying chemoradiation sensitive genotypes (GT/TT) had a better OS than chemoradiation resistant genotypes (GG) (median survival time [MST] 30 vs. 20 months [mo], hazard ratio [HR] = 0.628, 95% confidence interval [CI] 0.420-0.940, P=0.023). In contrast, among those treated with high dose RT, patients with chemoradiation sensitive genotypes (GT/TT) had a numerically worse OS than those with chemoradiation resistant genotypes (GG) (MST 22 vs. 31 mo, HR=1.57, 95% CI 0.86-2.83, P=0.13). Similar results were also found among patients received concurrent chemoradiation without cetuximab. This study validated our previous findings that ERCC2: rs238406 is a predictive biomarker of radiation dose with respect to OS in patients with NSCLC treated with concurrent chemoradiotherapy. Dose escalation may have worse survival in patients carrying chemoradiation sensitive genotypes. Genetic variants and their interaction with radiation dose may potentially guide individualized dose escalation in NSCLC radiotherapy.

PO-1544: Comparing Clinical Variables and Quantitative Imaging Features for Lung Cancer Survival Prediction

PO-1544: Comparing Clinical Variables and Quantitative Imaging Features for Lung Cancer Survival Prediction

The major effects of health-related quality of life on 5-year survival prediction among lung cancer survivors: applications of machine learning

The primary goal of this study was to evaluate the major roles of health-related quality of life (HRQOL) in a 5-year lung cancer survival prediction model using machine learning techniques (MLTs). The predictive performances of the models were compared with data from 809 survivors who underwent lung cancer surgery. Each of the modeling technique was applied to two feature sets: feature set 1 included clinical and sociodemographic variables, and feature set 2 added HRQOL factors to the variables from feature set 1. One of each developed prediction model was trained with the decision tree (DT), logistic regression (LR), bagging, random forest (RF), and adaptive boosting (AdaBoost) methods, and then, the best algorithm for modeling was determined. The models’ performances were compared using fivefold cross-validation. For feature set 1, there were no significant differences in model accuracies (ranging from 0.647 to 0.713). Among the models in feature set 2, the AdaBoost and RF models outperformed the other prognostic models [area under the curve (AUC) = 0.850, 0.898, 0.981, 0.966, and 0.949 for the DT, LR, bagging, RF and AdaBoost models, respectively] in the test set. Overall, 5-year disease-free lung cancer survival prediction models with MLTs that included HRQOL as well as clinical variables improved predictive performance.

FOXM1 and UBE2C Are Distinct Biomarkers for Non-Small Cell Lung Cancer Survival Prediction: Data-Mining Based on ONCOMINE

Non-small cell lung cancer (NSCLC) remains to be primary reason of tumor deaths in the past few decades. The mortality of this malignancy could be reduced by developing new prognostic biomarkers and discovering novel therapeutic biological target. Here, we studied the mRNA expression of FOX gene family and UBE2C in different types of cancer compared with normal tissue through ONCOMINE differential analysis. CCLE analysis was mined to explore the expression profiles of target genes in different tumor cells. GEPIA was used to discover the expression of target genes in different subtypes and the correlation with lung cancer stage. The prognostic values of FOXM1 and UBE2C were further investigated through Kaplan-Meier plotter analysis. It showed that FOXA1, FOXD1 and FOXM1 were dramatically high expressed in NSCLC comparing with normal lung tissues. Besides, the expression of FOXM1 was significantly associated with UBE2C. Furthermore, the overexpression of FOXM1 and UBE2C were correlated to shorter survival in lung adenocarcinoma (LAC) instead of lung squamous cell carcinoma(LSCC).Hence, we could draw a conclusion that FOXM1 and UBE2C are distinguished biomarkers and crucial prognostic indicators for lung adenocarcinoma patients.

Lung Cancer Survival Prediction via Machine Learning Regression, Classification, and Statistical Techniques.

A regression model is developed to predict survival time in months for lung cancer patients. It was previously shown that predictive models perform accurately for short survival times of less than 6 months; however, model accuracy is reduced when attempting to predict longer survival times. This study employs an approach for which regression models are used in combination with a classification model to predict survival time. A set of de-identified lung cancer patient data was obtained from the Surveillance, Epidemiology, and End Results (SEER) database. The models use a subset of factors selected by ANOVA. Model accuracy is measured by a confusion matrix for classification and by Root Mean Square Error (RMSE) for regression. Random Forests are used for classification, while general Linear Regression, Gradient Boosted Machines (GBM), and Random Forests are used for regression. The regression results show that RF had the best performance for survival times ≤6 and >24 months (RMSE 10.52 and 20.51, respectively), while GBM performed best for 7-24 months (RMSE 15.65). Comparison plots of the results further indicate that the regression models perform better for shorter survival times than the RMSE values are able to reflect.

The effect of imputing missing clinical attribute values on training lung cancer survival prediction model performance.

According to the estimations of the World Health Organization and the International Agency for Research in Cancer, lung cancer is the most common cause of death from cancer worldwide. The last few years have witnessed a rise in the attention given to the use of clinical decision support systems in medicine generally and in cancer in particular. These can predict patients' likelihood of survival based on analysis of and learning from previously treated patients. The datasets that are mined for developing clinical decision support functionality are often incomplete, which adversely impacts the quality of the models developed and the decision support offered. Imputing missing data using a statistical analysis approach is a common method to addressing the missing data problem. This work investigates the effect of imputation methods for missing data in preparing a training dataset for a Non-Small Cell Lung Cancer survival prediction model using several machine learning algorithms. The investigation includes an assessment of the effect of imputation algorithm error on performance prediction and also a comparison between using a smaller complete real dataset or a larger dataset with imputed data. Our results show that even when the proportion of records with some missing data is very high (>80%) imputation can lead to prediction models with an AUC (0.68-0.72) comparable to those trained with complete data records.