Integrated imaging and molecular analysis to decipher tumor microenvironment in the era of immunotherapy

Abstract

Radiological imaging is an integral component of cancer care, including diagnosis, staging, and treatment response monitoring. It contains rich information about tumor phenotypes that are governed not only by cancer cellintrinsic biological processes but also by the tumor microenvironment, such as the composition and function of tumor-infiltrating immune cells. By analyzing the radiological scans using a quantitative radiomics approach, robust relations between specific imaging and molecular phenotypes can be established. Indeed, a number of studies have demonstrated the feasibility of radiogenomics for predicting intrinsic molecular subtypes and gene expression signatures in breast cancer based on MRI. In parallel, promising results have been shown for inferring the amount of tumor-infiltrating lymphocytes, a key factor for the efficacy of cancer immunotherapy, from standard-of-care radiological images. Compared with the biopsy-based approach, radiogenomics offers a unique avenue to profile the molecular makeup of the tumor and immune microenvironment as well as its evolution in a noninvasive and holistic manner through longitudinal imaging scans. Here, we provide a systematic review of the state of the art radiogenomics studies in the era of immunotherapy and discuss emerging paradigms and opportunities in AI and deep learning approaches. These technical advances are expected to transform the radiogenomics field, leading to the discovery of reliable imaging biomarkers. This will pave the way for their clinical translation to guide precision cancer therapy.

Keywords: Immunotherapy; Machine learning; Radiogenomics; Radiomics; Tumor immune microenvironment.

Figure 1.

Four tasks of implementing artificial intelligence for noninvasively analyzing radiological scans in cancer patient management during the course of cancer development, where task 1 focuses on improving the accuracy of cancer diagnosis; task 2 focuses on tailoring individual treatment planning; task 3 focuses on monitoring and following up treatment response; task 4 focuses on early abnormality detection.

Figure 2.

The computational pipeline of key steps involved in a standard radiomic analysis, including patient enrollment, imaging preprocessing, feature extraction, as well as predictive model training and validation.

Figure 3.

Illustration of standard radiogenomic analysis, which encompasses twofold. One direction is to build radiological surrogates for established molecular biomarkers, which can serve as a way of virtual biopsy to noninvasively monitor underlying biological patterns as well as tumor response to specific therapy; another direction is to flip the analysis and decipher the biological mechanism behind putative imaging findings.

Figure 4.

Forest plot of reported sensitivity (A) and specificity (B) in radiogenomic studies using radiomic features from dynamic contrast-enhancing MRI for molecular subtype prediction in breast cancer; A meta-analysis was carried with R package mada to summarize the reported studies with S-ROC curve.

Figure 5.

A) Illustration of three breast cancer subtypes defined by the radiomic features extracted from the primary tumor and background parenchyma; B) Kaplan-Meier curve of recurrence-free survival stratified by predicted imaging subtype through the radiogenomic classifier of gene expression data from GEO.

Figure 6.

MR radiomic features were reported to be correlated with TIL levels in breast cancer. The bell curve represents the statistical significance of the correlation, where the center part represents a higher confidence level. The radiomic features are arranged based on the direction of correlation and colored with a blue circle if associated with immune suppression and with a red circle for immunogenic effects.

Figure 7.

The flowchart of the study design to integrate radiologic, pathologic, and genomic information to comprehensively investigate their correlation regarding tumor-infiltrating lymphocytes (TILs). Three key steps were planned. Reprinted with permission from Jia Wu, Stanford University (Wu, J., Li, X., Teng, X., Rubin, D. L., Napel, S., Daniel, B. L., & Li, R. (2018). Magnetic resonance imaging and molecular features associated with tumor-infiltrating lymphocytes in breast cancer. Breast Cancer Research, 20(1), 101).

Figure 8.

Illustration of three breast cancer patients in the TCGA cohort, where the radiomic model can accurately predict their tumor-infiltrating lymphocyte (TIL) value. Reprinted with permission from Jia Wu, Stanford University (Wu, J., Li, X., Teng, X., Rubin, D. L., Napel, S., Daniel, B. L., & Li, R. (2018). Magnetic resonance imaging and molecular features associated with tumor-infiltrating lymphocytes in breast cancer. Breast Cancer Research, 20(1), 101).

Figure 9.

The overview of three main computational algorithms to define habitats based on radiological scans. The first one is to filter the tumor with a predefined cutoff value; the second one is to partition the tumor into intratumor versus peritumor regions; the third one is a clustering-based analysis, which can uncover the intrinsic habitats within the tumor as well as its interactions with tumor-adjacent parenchyma.

Figure 10.

The blueprint of integrating radiological scans with biospecimen-related measurement for augmenting cancer patient management, where both challenges and opportunities exist to translate pioneer findings into clinical practice.