Abstract: Evolution and variation of tumor morphology during treatment may reflect or correlate with outcomes and contribute toward adapted response criteria.
Introduction
Standard radiological criteria, specifically RECIST 1.1, have long served as the benchmark for assessing cancer treatment efficacy. However, the advent of immune checkpoint inhibitors (ICIs) has challenged this paradigm. Immunotherapy often induces atypical response patterns, such as pseudoprogression—where a tumor appears to enlarge due to immune cell infiltration—or dissociated responses, where different lesions within the same patient react differently. In these scenarios, size-based metrics may prematurely categorize a patient as having progressive disease, potentially leading to the cessation of a beneficial treatment. This study explores the utility of tumor morphology, specifically "solidity," as a more nuanced prognostic biomarker.
Methods
This retrospective study analyzed 37 patients with unresectable primary liver cancer (hepatocellular carcinoma or biliary cancer) enrolled in clinical trials at the National Institutes of Health. Patients received a combination of immunotherapy (tremelimumab ± durvalumab) and locoregional treatments like ablation or chemoembolization.
Radiologists performed manual segmentation of the largest target lesions on 2D axial CT or MRI slices at baseline, 3-month, and 6-month intervals. Using ImageJ, two primary morphological descriptors were calculated:
- Solidity: The ratio of the tumor area to the area of its "convex hull" (the smallest convex polygon that encloses it). This metric quantifies the regularity or "jaggedness" of the tumor border.
- Circularity: A measure of how closely the tumor's shape approximates a perfect circle.
Key Insights and Ideas
The study found that morphological features provide information independent of traditional size-based measurements. Interestingly, shape features did not correlate with RECIST 1.1 status, suggesting that morphology captures biological changes that diameter-based metrics overlook.
The most significant finding was the prognostic power of solidity. While baseline morphology was not predictive of survival, the evolution of these shapes during treatment was highly telling. Specifically, a decrease in solidity (indicating a more irregular, invasive-looking border) by the 6-month mark was a strong indicator of poor prognosis.
Key Results
Statistical analysis revealed that patients with low-solidity tumors at 6 months had significantly poorer overall survival ($p=0.01$). The researchers identified an optimal solidity threshold of 0.956; patients exceeding this threshold had a median survival of 525 days, compared to just 304 days for those below it.
Furthermore, a decrease in tumor solidity between baseline and 6 months was also strongly associated with poorer survival ($p=0.01$). While circularity showed variation over time ($p=0.04$), it was not as robust a predictor of overall survival as solidity.
Discussion and Limitations
While promising, this study has several limitations. The use of 2D manual segmentation, while practical for current workflows, may be less comprehensive than 3D volumetric analysis. The sample size of 37 patients is relatively small and heterogeneous, mixing HCC and biliary cancers.
Future research should focus on automating these measurements using deep-learning tools to reduce labor and inter-observer variability. Additionally, integrating shape descriptors with other radiomic features, such as enhancement patterns, could lead to more robust, multi-parametric response criteria tailored for the immunotherapy era.
Summary updated