Design and Implementation of an Explainable Decision-Support Pipeline for Glioma  Grading and Brain Tumor MRI Classification

Kavya K; Mr. S. Manjunatha

Design and Implementation of an Explainable Decision-Support Pipeline for Glioma Grading and Brain Tumor MRI Classification

Author(s)	Ms. Kavya K, Mr. S. Manjunatha
Country	India
Abstract	—This paper documents the system design and implementation of a two-pipeline clinical decision-support tool for neuro-oncology: a tabular classifier that grades glioma severity from TCGA mutation and clinical records, and a convolutional image classifier that sorts brain MRI scans into four diagnostic categories. Rather than treating predictive accuracy as the only deliverable, we walk through the engineering pipeline that turns trained models into an auditable tool — data ingestion and cleaning, class-imbalance correction, model selection, explanation generation, and web-based deployment. On the tabular side, eight classifier families are trained on 23 engineered features drawn from 839 TCGA patient records; Logistic Regression reaches 86.90% held-out accuracy, and Random Forest attains the top AUC-ROC of 0.9229 (5-fold CV AUC 0.9258). A SHAP TreeExplainer module attached to the XGBoost model produces both global and per-patient feature attributions, surfacing IDH1 mutation status as the dominant predictor, in line with WHO 2021 grading criteria. On the imaging side, ResNet50 and MobileNetV2 are fine-tuned through transfer learning for four-class MRI classification (glioma, meningioma, pituitary, no tumor), reaching up to 95.58% validation accuracy and combined at inference time via softvote averaging. Both pipelines sit behind a single Flask web application offering an interactive dashboard for tabular prediction, MRI upload, and real-time SHAP visualization. We describe the module-level architecture, the training and inference workflows, and the evaluation protocol used to validate the system, and close with deployment considerations and current limitations.
Keywords	—Clinical decision-support systems, explainable AI, SHAP, glioma grading, TCGA, transfer learning, ResNet50, MobileNetV2, brain tumor MRI classification, system deployment
Field	Computer > Artificial Intelligence / Simulation / Virtual Reality
Published In	Volume 7, Issue 4, July-August 2026
Published On	2026-08-05

View / Download PDF File

E-ISSN 3048-7641

doi

CrossRef DOI is assigned to each research paper published in our journal.

AIJFR DOI prefix is
10.63363/aijfr

Downloads

Research Paper Format Copyright Permission Form and Undertaking Form Cover Page Vol 7 Isu 4 Cover Page Vol 7 Isu 3 Cover Page Vol 7 Isu 2

All research papers published on this website are licensed under Creative Commons Attribution-ShareAlike 4.0 International License, and all rights belong to their respective authors/researchers.

CC-BY-SA

About AIJFR Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics Peer Review & Plagiarism	Join as a Reviewer Editors & Reviewers Reviewer Referral Program Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us		+91-9898-79-39-59	editor@aijfr.com

Advanced International Journal for Research

A Widely Indexed Open Access Peer Reviewed Multidisciplinary Bi-monthly Scholarly International Journal

Design and Implementation of an Explainable Decision-Support Pipeline for Glioma Grading and Brain Tumor MRI Classification

Share this