Study: AI May Help Predict Response to Immune Checkpoint Blockade in Patients With Metastatic Melanoma

To analyze digital images of metastatic tumors of melanoma, the researchers used computer algorithms, or deep convolutional neural networks (DCCN), to identify patterns associated with treatment response.

A computational method that combines clinicodemographic variables with deep learning of pre-treatment histology images could predict response to immune checkpoint blockade among patients with advanced melanoma, according to results published in Clinical Cancer Research.

“While immune checkpoint inhibitors have profoundly changed the treatment landscape in melanoma, many tumors do not respond to treatment, and many patients experience treatment-related toxicity,” said corresponding study author Iman Osman, MD, medical oncologist in the Departments of Dermatology and Medicine at New York University (NYU) Grossman School of Medicine, in a press release. “An unmet need is the ability to accurately predict which tumors will respond to which therapy. This would enable personalized treatment strategies that maximize the potential for clinical benefit and minimize exposure to unnecessary toxicity.”

The researchers used data from a training cohort of 121 patients with metastatic melanoma who received immune checkpoint blockade treatment between 2004 and 2018. All of the patients were treated with first-line anti-CTLA-4 therapy, anti-programmed cell death protein 1 (PD-1) therapy, or a combination of both or partial responses. The study authors noted that patients with stable disease were excluded for this proof-of-principle study.

To analyze digital images of metastatic tumors of melanoma, the researchers used computer algorithms, or deep convolutional neural networks (DCCN), to identify patterns associated with treatment response. This approach helped the team develop a response classifier, which aimed to predict whether a patient’s untreated tumor would respond to immune checkpoint blockade or progress following treatment. Further, the DCCN response classifier was validated in an independent cohort of 30 patients with metastatic melanoma treated at Vanderbilt-Ingram Cancer Center between 2010 and 2017, according to the study.

The performance of the DCCN response classifier was evaluated by calculating the area under the curve (AUC), which is the measure of the model’s accuracy, in which a score of 1 corresponds to perfect prediction. Additionally, the DCCN prediction model achieved an AUC around 0.7 in both the training and validation cohorts.

Further, the researchers performed multivariable logistic regressions that combined the DCCN prediction with conventional clinical characteristics to augment the prediction accuracy of the model. The final model incorporated the DCCN prediction, Eastern Cooperative Oncology Group performance status, and treatment regimen (either anti-CTLA-4 monotherapy, anti PD-1 monotherapy, or combination therapy), according to the study.

In both the training and validation cohorts, the multivariable classifier achieved an AUC around 0.8. In the validation cohort, the classifier could stratify patients into high- versus low-risk for disease progression, with significantly different progression-free survival outcomes between the 2 groups. Whereas 64% of patients in the training cohort received anti-CTLA-4 monotherapy, approximately 53% of the patients in the validation cohort received anti-PD-1 agents.

The results of the study suggest that some predictive patterns are not specific to the immune checkpoint target, according to the study authors. Additionally, class activation mapping, which can help identify regions within the digital images that the neural network uses to generate predictions, suggests that cell nuclei were important for the DCCN predictions, according to the researchers.

“There is potential for using computer algorithms to analyze histology images and predict treatment response, but more work needs to be done using larger training and testing datasets, along with additional validation parameters, in order to determine whether an algorithm can be developed that achieves clinical-grade performance and is broadly generalizable,” said study author Aristotelis Tsirigos, PhD, professor in the Institute for Computational Medicine at NYU Grossman School of Medicine, in a press release.

Study limitations included the relatively small number of images used to train the computer algorithm, which included 302 images in the training cohort and 40 images in the validation cohort.

REFERENCE

Artificial intelligence may help predict response to immune checkpoint blockade in patients with metastatic melanoma. American Association of Cancer Research. Published November 18, 2020. Accessed November 18, 2020. https://www.aacr.org/about-the-aacr/newsroom/news-releases/artificial-intelligence-may-help-predict-response-to-immune-checkpoint-blockade-in-patients-with-metastatic-melanoma/.