Dataset Information

A deep learning-based system for survival benefit prediction of tyrosine kinase inhibitors and immune checkpoint inhibitors in stage IV non-small cell lung cancer patients: A multicenter, prognostic study.

ABSTRACT:

Background

For clinical decision making, it is crucial to identify patients with stage IV non-small cell lung cancer (NSCLC) who may benefit from tyrosine kinase inhibitors (TKIs) and immune checkpoint inhibitors (ICIs). In this study, a deep learning-based system was designed and validated using pre-therapy computed tomography (CT) images to predict the survival benefits of EGFR-TKIs and ICIs in stage IV NSCLC patients.

Methods

This retrospective study collected data from 570 patients with stage IV EGFR-mutant NSCLC treated with EGFR-TKIs at five institutions between 2010 and 2021 (data of 314 patients were from a previously registered study), and 129 patients with stage IV NSCLC treated with ICIs at three institutions between 2017 and 2021 to build the ICI test dataset. Five-fold cross-validation was applied to divide the EGFR-TKI-treated patients from four institutions into training and internal validation datasets randomly in a ratio of 80%:20%, and the data from another institution was used as an external test dataset. An EfficientNetV2-based survival benefit prognosis (ESBP) system was developed with pre-therapy CT images as the input and the probability score as the output to identify which patients would receive additional survival benefit longer than the median PFS. Its prognostic performance was validated on the ICI test dataset. For diagnosing which patient would receive additional survival benefit, the accuracy of ESBP was compared with the estimations of three radiologists and three oncologists with varying degrees of expertise (two, five, and ten years). Improvements in the clinicians' diagnostic accuracy with ESBP assistance were then quantified.

Findings

ESBP achieved positive predictive values of 80·40%, 75·40%, and 77·43% for additional EGFR-TKI survival benefit prediction using the probability score of 0·2 as the threshold on the training, internal validation, and external test datasets, respectively. The higher ESBP score (>0·2) indicated a better prognosis for progression-free survival (hazard ratio: 0·36, 95% CI: 0·19-0·68, p<0·0001) in patients on the external test dataset. Patients with scores >0·2 in the ICI test dataset also showed better survival benefit (hazard ratio: 0·33, 95% CI: 0·18-0·55, p<0·0001). This suggests the potential of ESBP to identify the two subgroups of benefiting patients by decoding the commonalities from pre-therapy CT images (stage IV EGFR-mutant NSCLC patients receiving additional survival benefit from EGFR-TKIs and stage IV NSCLC patients receiving additional survival benefit from ICIs). ESBP assistance improved the diagnostic accuracy of the clinicians with two years of experience from 47·91% to 66·32%, and the clinicians with five years of experience from 53·12% to 61·41%.

Interpretation

This study developed and externally validated a preoperative CT image-based deep learning model to predict the survival benefits of EGFR-TKI and ICI therapies in stage IV NSCLC patients, which will facilitate optimized and individualized treatment strategies.

Funding

This study received funding from the National Natural Science Foundation of China (82001904, 81930053, and 62027901), and Key-Area Research and Development Program of Guangdong Province (2021B0101420005).

SUBMITTER: Deng K

PROVIDER: S-EPMC9256845 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:Immune checkpoint inhibitors (ICIs) have been approved to treat patients with various cancer types, including lung cancer, in many countries. This study aims to investigate the effectiveness and safety of ICIs under different treatment conditions of non-small cell lung cancer patients. A population-based retrospective cohort study was conducted using the electronic health records of three medical centers in Taiwan. From January 01, 2016, to November 30, 2018, a total of 91 ICIs and 300 traditional chemotherapy users who had undergone stage III and IV lung cancer treatment were included in the study. We performed the randomized matched pair design by selecting a Chemotherapy subject for each ICI patient in the sample population. All subjects were monitored from the date of taking ICIs or chemotherapy drugs until the event of death, loss to follow-up, or were occurred with any defined adverse events. Kaplan-Meier estimators and cox proportional hazard regression models were used to compute the overall survival, efficacy, and safety of the ICIs group. The median overall survival (OS) in the ICI and Chemo groups after matching was 11.2 months and 10.5 months, respectively. However, the results showed no significant OS differences between ICIs and chemo groups for both before and after matching (HR,1.30; 95%CI, 0.68-2.46; p=0.428 before matching and HR,0.96; 95CI%, 0.64-1.44; p=0.838 after matching). We observed that with the higher amount of PD-L1, the length of the patients' overall survival was (positive vs. negative PD-L1, HR,0.21; 95%CI, 0.05-0.80; p=0.022). The incidences of serious adverse drug events above grade 3 in the ICIs and traditional chemo groups were 12.7% and 21.5%, respectively. We also found that the number of AEs was less in ICIs than in the Chemo group, and the AEs that occurred after treatments were observed earlier in the ICIs compared to the Chemo group. ICIs drugs were observed to be safer than traditional chemotherapy as they had a lower risk of serious adverse drug events. It is necessary to pay attention to immune-related side effects and provide appropriate treatment. Furthermore, the patient's physical status and PD-L1 test can be used to evaluate the clinical effectiveness of ICIs.

Project description:BackgroundImmunotherapy with immune checkpoint inhibitors (ICIs) is highly effective in microsatellite instability-high (MSI-H) metastatic colorectal cancer (mCRC); however, specific predictive biomarkers are lacking.Patients and methodsData and samples from 85 patients with MSI-H mCRC treated with ICIs were gathered. Tumor infiltrating lymphocytes (TILs) and tumor mutational burden (TMB) were analyzed in an exploratory cohort of "super" responders and "clearly" refractory patients; TILs were then evaluated in the whole cohort of patients. Primary objectives were the correlation between the number of TILs and TMB and their role as biomarkers of ICI efficacy. Main endpoints included response rate (RR), progression-free survival (PFS), and overall survival (OS).ResultsIn the exploratory cohort, an increasing number of TILs correlated to higher TMB (Pearson's test, p = .0429). In the whole cohort, median number of TILs was 3.6 in responders compared with 1.8 in nonresponders (Mann-Whitney test, p = .0448). RR was 70.6% in patients with high number of TILs (TILs-H) compared with 42.9% in patients with low number of TILs (odds ratio = 3.20, p = .0291). Survival outcomes differed significantly in favor of TILs-H (PFS: hazard ratio [HR] =?0.42, p = .0278; OS: HR = 0.41, p = .0463).ConclusionA significant correlation between higher TMB and increased number of TILs was shown. A significantly higher activity and better PFS and OS with ICI in MSI-H mCRC were reported in cases with high number of TILs, thus supporting further studies of TIL count as predictive biomarker of ICI efficacy.Implications for practiceMicrosatellite instability is the result of mismatch repair protein deficiency, caused by germline mutations or somatic modifications in mismatch repair genes. In metastatic colorectal cancer (mCRC), immunotherapy (with immune checkpoint inhibitors [ICIs]) demonstrated remarkable clinical benefit in microsatellite instability-high (MSI-H) patients. ICI primary resistance has been observed in approximately 25% of patients with MSI-H mCRC, underlining the need for predictive biomarkers. In this study, tumor mutational burden (TMB) and tumor infiltrating lymphocyte (TIL) analyses were performed in an exploratory cohort of patients with MSI-H mCRC treated with ICIs, demonstrating a significant correlation between higher TMB and increased number of TILs. Results also demonstrated a significant correlation between high number of TILs and clinical responses and survival benefit in a large data set of patients with MSI-H mCRC treated with ICI. TMB and TILs could represent predictive biomarkers of ICI efficacy in MSI-H mCRC and should be incorporated in future trials testing checkpoint inhibitors in colorectal cancer.

Project description:PURPOSE:Locoregional treatment has been increasingly adopted for metastatic breast cancer at presentation. This study aims to develop an individualized calculator to predict the benefit of postoperative radiotherapy (PORT) for patients with surgically resected de novo stage IV breast cancer. METHODS AND MATERIALS:We searched the Surveillance, Epidemiology, and End Results (SEER) database for patients diagnosed with stage IV breast cancer between 2010 and 2014. After applying exclusion criteria, a total of 4473 patients were included in the analysis. Propensity score matching was used to balance the individual characteristics of the patients. After identifying the significant prognosticators, a nomogram was developed using multivariate regression models and internally validated. A web-based calculator was then constructed using a fitted survival prediction model. RESULTS:With a median follow-up of 34 months, the three-year overall survival (OS) rates were 54.1% in the surgery alone group and 63.5% in the surgery + PORT group (p < 0.001). The survival benefit of PORT was maintained after propensity score matching (p < 0.001). Interaction testing of the prognostic variables found significant interactions between PORT and the presence of brain metastasis (p = 0.001), and between PORT and hormonal receptor expression (p = 0.018). After reviewing the performance of various models, a log-normal distributed survival model was adopted, with a C-index of 0.695. A calibration plot verified that the predicted survival rates were strongly correlated with the actual OS rates. A web-based survival calculator was constructed to provide individualized estimates of survival according to PORT. CONCLUSION:PORT significantly improved OS rates, though the individual benefit was affected by a number of factors. We successfully developed a nomogram and web-based calculator that predicted the prognosis according to PORT in patients with surgically resected de novo stage IV breast cancer. These tools are expected to be useful in clinical practice and in the design of related trials.

Project description:BACKGROUND The early death of patients is a global cancer issue. We aimed to identify the risk factors for early death in stage IV breast cancer. Predictive nomograms for early death evaluation were generated based on the risk factors. MATERIAL AND METHODS Based on the Surveillance, Epidemiology, and End Results (SEER) database, patients diagnosed with IV breast cancer were selected. The risk factors for early death (survival time ?1 year) were identified using logistic regression model analysis. Predictive nomograms were constructed and internal validation was performed. RESULTS A total of 5998 (32.6%) breast cancer patients were diagnosed as early death in the construction cohort. Age older than 50 years, unmarried status, black race, uninsured status, triple-negative type, grade (II and III), tumor size >5 cm, and metastasis to lung, liver, and brain were risk factors for total early death, while Luminal B subtype, N1 stage, and surgical interventions were associated with lower risk of early death. As for cancer-specific and non-cancer-specific early death, several factors were not consistent between the 2 groups. Nomograms for all-cause, cancer-specific, and non-cancer-specific early death were constructed. The calibration curve showed satisfactory agreement. The areas under the ROC curve (AUC) were 78.3% (95% CI: 77.7-78.9%), 75.8% (75.1-76.4%), and 72.3% (71.6-72.9%), respectively. In the validation cohort, a total of 689 (19.3%) patients were diagnosed as early death and the calibration curve showed satisfactory agreement. The AUCs of the all-cause, cancer-specific, and non-cancer-specific early death prediction were 74.0% (95% CI: 72.5-75.4%), 73.5% (72.0-74.9%), and 68.6% (67.0-70.1%), respectively. CONCLUSIONS Nomograms were generated to predict early death, with good calibration and discrimination. The predictive model can provide a reference for identifying cases with high risk of early death among stage IV breast cancer patients and play an auxiliary role in guiding individual treatment.