Harnessing neural networks to forecast treatment outcomes in small cell lung cancer

FrontiersOct 12 2024

As immunotherapy makes breakthrough progress in the treatment of small cell lung cancer (SCLC), predicting treatment outcomes has become a focal point in clinical practice. Immunotherapy combined with chemotherapy has been approved as first-line therapy for small cell lung cancer due to its survival benefit in randomized controlled trials.

However, predicting its efficacy remains a challenge in the absence of currently available biomarkers. Recently, a study titled "Neural network models based on clinical characteristics for predicting immunotherapy efficacy in small Cell lung cancer" was published in Malignancy Spectrum. The study utilized deep learning techniques to develop a novel predictive model, providing clinicians with a powerful decision-making aid.
The research team retrospectively analyzed data from 140 SCLC patients who underwent immunotherapy, dividing them into a discovery cohort and a validation cohort. By constructing and training neural network models, predictive models for three clinical outcomes were developed, and the researchers successfully predicted the objective response rate (ORR), disease control rate (DCR), and the proportion of patients with progression-free survival (PFS) over six months. The study results showed that the ORR model achieved an AUC value of 0.8964 in the discovery cohort and 0.8421 in the validation cohort, demonstrating high predictive accuracy (Figure 1). The models were then compressed into a doctor-friendly tool.
This research not only provides new scientific evidence for personalized treatment of SCLC patients but also offers an important reference for future clinical decisions regarding immunotherapy. The research team stated that they will continue to optimize the model and further validate its stability and universality in prospective studies with multicenter and large samples.
This study was conducted by the Cancer Research Center at Beijing Chest Hospital, Capital Medical University/Beijing Tuberculosis and Thoracic Tumor Research Institute, with contributions from Sichuan Provincial People's Hospital, University of Electronic Science and Technology of China, and the Department of Medical Oncology at Beijing Tuberculosis and Thoracic Tumor Research Institute/Beijing Chest Hospital, Capital Medical University.