Back to all available

Effectiveness of Machine Learning in Identifying Lymphovascular and Perineural Invasion in Gastric Cancer: A Systematic Review Meta-Analysis

evidence-based-medicine
oncology
pathology
radiology-and-medical-imaging
data-mining-and-machine-learning

Abstract

Background. For gastric cancer(GC), accurately identifying lymphovascular invasion (LVI) and perineural invasion (PNI) is critical for guiding treatment decisions. In recent years, researchers have explored machine learning (ML) for preoperative detection of LVI and PNI in GC; however, systematic evidence on its diagnostic accuracy remains scarce. To address this gap, this study was undertaken to comprehensively review ML's performance for detecting LVI and PNI, thereby providing empirical evidence to enhance or update intelligent diagnostic techniques.

Methods. Multiple databases, PubMed, the Cochrane Library, Embase, and Web of Science, were retrieved for ML articles on the prediction of LVI and PNI in GC, covering publications from inception through January 13, 2025. Researchers employed PROBAST to ascertain the potential risk of bias in the selected studies. Subgroup analyses were executed by modeling variables. All meta-analyses were executed in Stata 15.0.

Results. Twenty-four eligible studies were ultimately incorporated, most of which employed logistic regression models (LRMs). For LVI diagnosis in the validation sets, models using clinical features (CFs) alone demonstrated a sensitivity (SEN) of 0.71 (95% CI: 0.61-0.80) and a specificity (SPC) of 0.77 (95% CI: 0.67-0.85). Models utilizing radiomics features (RFs) alone yielded a SEN of 0.80 (95% CI: 0.71-0.86) and a SPC of 0.66 (95% CI: 0.56-0.75). The models combining CFs and RFs outperformed others, with a SEN of 0.81 (95% CI: 0.75-0.86) and a SPC of 0.73 (95% CI: 0.65-0.80). Regarding the detection of PNI in validation sets, CF-based models yielded a SEN of 0.71 (95% CI: 0.60-0.80) with a corresponding SPC of 0.62 (95% CI: 0.52-0.71). RF-based models demonstrated a SEN of 0.75 (95% CI: 0.71-0.79) and a SPC of 0.76 (95% CI: 0.72-0.79); while the models based on both CFs and RFs exhibited superior performance with a SEN of 0.81 (95% CI: 0.76-0.84) and a SPC of 0.82 (95% CI: 0.79-0.85).

Conclusions. ML models combining radiomics with CFs appear to be a feasible approach for identifying LVI and PNI in GC. However, the number of included studies was small, and these studies used a random split approach for internal validation. Future research should incorporate more multicenter studies to develop more robust prediction models for clinical implementation.

Ask to review this manuscript

Notes for potential reviewers

  • Volunteering is not a guarantee that you will be asked to review. There are many reasons: reviewers must be qualified, there should be no conflicts of interest, a minimum of two reviewers have already accepted an invitation, etc.
  • This is NOT OPEN peer review. The review is single-blind, and all recommendations are sent privately to the Academic Editor handling the manuscript. All reviews are published and reviewers can choose to sign their reviews.
  • What happens after volunteering? It may be a few days before you receive an invitation to review with further instructions. You will need to accept the invitation to then become an official referee for the manuscript. If you do not receive an invitation it is for one of many possible reasons as noted above.
  • PeerJ does not judge submissions based on subjective measures such as novelty, impact or degree of advance. Effectively, reviewers are asked to comment on whether or not the submission is scientifically and technically sound and therefore deserves to join the scientific literature. Our Peer Review criteria can be found on the "Editorial Criteria" page - reviewers are specifically asked to comment on 3 broad areas: "Basic Reporting", "Experimental Design" and "Validity of the Findings".
  • Reviewers are expected to comment in a timely, professional, and constructive manner.
  • Until the article is published, reviewers must regard all information relating to the submission as strictly confidential.
  • When submitting a review, reviewers are given the option to "sign" their review (i.e. to associate their name with their comments). Otherwise, all review comments remain anonymous.
  • All reviews of published articles are published. This includes manuscript files, peer review comments, author rebuttals and revised materials.
  • Each time a decision is made by the Academic Editor, each reviewer will receive a copy of the Decision Letter (which will include the comments of all reviewers).

If you have any questions about submitting your review, please email us at [email protected].