AI identifies patterns on CT scans that offer new promise for treating small cell lung cancer

Researchers on the Heart for Computational Imaging and Personalised Diagnostics (CCIPD) at Case Western Reserve College have used synthetic intelligence (AI) to determine patterns on computed tomography (CT) scans that supply new promise for treating sufferers with small cell lung most cancers.

Small cell lung most cancers (SCLC) represents about 13% of all lung cancers, however grows sooner and is extra more likely to unfold than non-small cell lung most cancers, in accordance with the American Most cancers Society.

And whereas a variety of AI analysis has been carried out on non-small cell lung most cancers, little work has been accomplished on SCLC, stated CCIPD Director Anant Madabhushi, the Donnell Institute Professor of Biomedical Engineering at Case Western Reserve.

Small cell lung most cancers sufferers could be difficult to deal with, Madabhushi stated. His lab labored with oncologists at College Hospitals in Cleveland to assist confirm which SCLC sufferers would reply to therapy.

The researchers recognized a set of radiomic patterns from CT scans taken earlier than therapy that permit them to foretell a affected person’s response to chemotherapy. In addition they examined the affiliation between AI-derived picture options with longer-term outcomes.

Particularly, the researchers famous that computationally extracted textural patterns of the tumor itself-;in addition to the area surrounding it-;have been discovered to be completely different in SCLC sufferers who responded nicely to a sure chemotherapy, in comparison with those that didn’t.

Additional, patterns have been revealed by the AI that corresponded to sufferers who ended up dwelling longer after therapy in comparison with those that didn’t.

Lastly, the AI revealed that there was notably extra heterogeneity, or variability, within the scanned pictures of sufferers who didn’t reply to chemo and had poorer probabilities of survival, Madabhushi stated.

What’s subsequent: attainable human trials

These findings from a retrospective examine now units the stage for potential AI pushed scientific trials for therapy administration of SCLC sufferers, Madabhushi stated.

Outcomes from the analysis have been revealed in Frontiers in Oncology in October.

Their findings are important as a result of chemotherapy stays the spine of systemic therapy, the researchers stated.

“Regardless that most sufferers reply to preliminary therapy, relapse is frequent and a subset of sufferers are chemo-resistant,” stated Prantesh Jain, co-lead creator of the examine whereas with the Division of Hematology and Oncology at College Hospitals. He is now an assistant professor of oncology at Roswell Park Complete Most cancers Heart in Buffalo.

Presently, there are not any clinically validated predictive biomarkers to pick out a subpopulation of sufferers with main chemoresistance or early recurrence.”

Prantesh Jain, co-lead creator of the examine

Broader AI initiative

The examine is a part of broader analysis performed at CCIPD to develop and apply novel AI and machine-learning approaches to diagnose and predict remedy responses for numerous ailments and indications of most cancers, together with breast, prostate, head and neck, mind, colorectal, gynecologic and pores and skin most cancers.

“Our efforts are geared toward lowering pointless chemotherapeutic remedies and thus lowering affected person struggling,” stated the examine’s co-lead creator Mohammadhadi Khorrami, a CCIPD researcher and PhD scholar in biomedical engineering at Case Western Reserve.

“By understanding which sufferers will profit from remedy, we will lower ineffective remedies and enhance extra aggressive remedy in sufferers who’ve suboptimal or no response to the first-line remedy.”


Case Western Reserve College

Journal reference:

Jain, P., et al. (2021) Novel Non-Invasive Radiomic Signature on CT Scans Predicts Response to Platinum-Based mostly Chemotherapy and Is Prognostic of Total Survival in Small Cell Lung Most cancers. Frontiers in Oncology.

Related posts Introduces Its Autonomous Customer Support Chatbot


Using AI to Detect Fractures on X-Rays Leads to Improved Diagnosis |


PayPal Taps AI/ML in Battle Against Fraudsters


Leave a Comment