A narrative review of prognosis prediction models for non-small cell lung cancer: what kind of predictors should be selected and how to improve models?

Abstract

Objective: To discover potential predictors and explore how to build better models by summarizing the existing prognostic prediction models of non-small cell lung cancer (NSCLC).Background: Research on clinical prediction models of NSCLC has experienced explosive growth in recent years. As more predictors of prognosis are discovered, the choice of predictors to build models is particularly important, and in the background of more applications of next-generation sequencing technology, gene-related predictors are widely used. As it is more convenient to obtain samples and followup data, the prognostic model is preferred by researchers.Methods: PubMed and the Cochrane Library were searched using the items "NSCLC", "prognostic model", "prognosis prediction", and "survival prediction" from 1 January 1980 to 5 May 2021. Reference lists from articles were reviewed and relevant articles were identified. Conclusions:The performance of gene-related models has not obviously improved. Relative to the innovation and diversity of predictors, it is more important to establish a highly stable model that is convenient for clinical application. Most of the prevalent models are highly biased and referring to PROBAST at the beginning of the study may be able to significantly control the bias. Existing models should be validated in a large external dataset to make a meaningful comparison.

Predictive risk score for isolated brain metastasis in non-small cell lung cancer

“…There are many studies that have been published on prediction models for NSCLC, however this study differs by utilizing risk factors to predict the probability of IBM at the time of diagnosis of NSCLC and guide if additional brain imaging is warranted. One review found numerous studies that have developed a risk score or nomogram used to predict survival, prognosis, or lymph node metastasis ( 22 ). One commonly used model is the World Health Organization Performance Status (WHO-PS) which is the Eastern Cooperative Oncology Group Performance Status (ECOG-PS) scale which is mainly used to assess prognosis; however, it is based on subjective measurements and can have wide variability between evaluators ( 22 ).…”

“…One review found numerous studies that have developed a risk score or nomogram used to predict survival, prognosis, or lymph node metastasis ( 22 ). One commonly used model is the World Health Organization Performance Status (WHO-PS) which is the Eastern Cooperative Oncology Group Performance Status (ECOG-PS) scale which is mainly used to assess prognosis; however, it is based on subjective measurements and can have wide variability between evaluators ( 22 ). To this researcher’s knowledge, this is the only study that has produced a risk score to predict synchronous IBM, which could aid clinicians in deciding if additional brain imaging is warranted in questionable cases.…”

Multiomics-Based Feature Extraction and Selection for the Prediction of Lung Cancer Survival

“…Since our goal is not to create the best classifier, we only transferred information on extracted molecular features between the datasets and not the model itself. For the same reason, we did not include any features derived from clinical data, like the TNM staging system [17], focusing on molecular data instead. For a complete list of TCGA and CTPAC-3 cases selected for this study, see Supplementary Table S1.…”

“…Early applications ranged from statistical decision theory [12] to neural networks [13], paving the way for more sophisticated algorithms like Bayesian networks [14], support vector machines [15], and deep learning methods [16]. All of these approaches are based on a variety of distinct data, ranging from medical images to molecular characteristics, some of which were reviewed for NSCLC in the context of prognosis prediction [17].…”

Construction of a predictive model for immunotherapy efficacy in lung squamous cell carcinoma based on the degree of tumor-infiltrating immune cells and molecular typing

“…A clinical prediction model is a tool that combines multiple predictors to evaluate the probability of an individual presenting with a certain disease or clinical outcome. Some clinical prediction models have potential value for screening, diagnosis, treatment, and prognostic prediction of lung cancer [15][16][17]. With the rapid development of high-throughput sequencing and bioinformatics analysis methods, obtaining cancer-related genomes, transcriptomes, and immune-related information has become readily easier.…”