Perfusion-weighted imaging with dynamic contrast enhancement (PWI/DCE) morphologic, qualitative, semiquantitative, and radiomics features predicting undifferentiated pleomorphic sarcoma (UPS) treatment response

Figures

• 

  Fig. 1. Summary of the methodology of this study. First, PWI/DCE images were acquired from UPS patients at baseline (BL), post-chemotherapy (PC), and post-radiation therapy (PRT). Second, tumors were manually contoured in three dimensions by an MD-trained research assistant, generating tumor volume of interest (VOI) segmentations. Third, features were extracted from the tumor VOIs, including morphologic patterns, time-intensity curves, semiquantitative parameters, and radiomics. Finally, statistical analyses were performed on the extracted features comparing responders vs. partial/non-responders.

• 

  Fig. 2. Scatter plots of RECIST, WHO, and volumetric measurements comparing responders vs. partial/non-responders across different time points in their treatment. BL: Baseline; PC: Post-Chemotherapy; PRT: Post-Radiation; PD: Progressive Disease; PR: Partial Response.

• 

  Fig. 3. Representative PWI/DCE images of six different UPS morphology patterns observed at PRT.

• 

  Fig. 4. PWI/DCE morphology patterns are classified by pathology-assessed treatment effect (PATE). Representative PWI/DCE morphology patterns from UPS patients with > 90% PATE (top panels), 31–89% PATE (middle panels) showing unipolar enhancement (red arrows), and 5–30% PATE (bottom panels) showing bipolar enhancement (red arrows).

• 

  Fig. 5. Post-radiation/pre-surgical morphology patterns from PWI/DCE organized by patient response. Note: All P-values resulted from one-way chi-squared tests.

• 

  Fig. 6. PWI/DCE time-intensity curve (TIC) types at baseline, post-chemotherapy, and post-radiation are organized by response. At post-radiation/pre-surgical, 100% of good responders demonstrated a TIC-type 2, 60% of partial responders expressed a TIC-type 5, and 83.3% of non-responders expressed a TIC-type 5. All p-values resulted from one-way chi-squared tests.

• 

  Fig. 7. Boxplots of seven semiquantitative perfusion kinetic parameters comparing responders vs. partial/non-responders at post-radiation (PRT). P-values resulted from two-tailed non-parametric Wilcoxon rank-sum tests. The wash-in rate (WiR) decreased 34% in responders relative to partial/non-responders, showing the highest statistical association with response. a.u.: Arbitrary Units.

• 

  Fig. 8. Comparison of seven PWI/DCE kinetic parameter maps from a responder (A) vs. a non-responder (B) at post-radiation (PRT). Table (C) summarizes the p-values resulting from two-tailed non-parametric Wilcoxon rank-sum tests comparing these parameters from the entire population of responders (n = 14) vs. partial/non-responders (n = 16) at PRT. TTP: Time to Peak; PE: Peak Enhancement; WiR: Wash-In Rate; WoR: Wash-Out Rate; WiAUC: Wash-In Area Under the Curve; WoAUC: Wash-Out Area Under the Curve; AUC: Area Under the Curve.

• 

  Fig. 9. Scatter plot of PWI/DCE wash-in rate (WiR) delta comparing responders vs. partial/non-responders across treatment. P-values resulted from two-tailed non-parametric Wilcoxon rank-sum tests. BL: Baseline; PC: Post-Chemotherapy; PRT: Post-Radiation.

• 

  Fig. 10. WiR map radiomics scatter plot (A) showing means with 95% confidence intervals and boxplots (B) of top 10 radiomic features comparing responders (R) vs. partial/non-responders (PR/NR) at post-radiation (PRT). All P-values resulted from two-tailed non-parametric Wilcoxon rank-sum tests. s.v.: Standardized Value.

• 

  Fig. 11. ROC curves for PWI/DCE capsular pattern, TIC type 2, and WiR, discriminating responders from partial/non-responders at PRT.

• 

  Fig. 12. Sarcoma organ model of undifferentiated pleomorphic sarcoma. On the left, a 3D CT-angiogram (CTA) reconstruction demonstrates a superior pole vascular supply dominance. A secondary, inferior vascular pole is also seen. On the right, a post-radiation PWI/DCE wash-in map in a UPS with < 90% TIN demonstrates arterial perfusion gradient from hypervascular residual solid components in proximity to arterial supply (superior pole/unipolar enhancement) to a central/equatorial necrosis in correlation with hypoxic tumor. The figure was modified with permission from the authors of RF Valenzuela et al.30.