1MO - Image biomarker discovery from DCE-MRI for identifying responders of MK-2206 on early-stage breast cancer patients: A secondary radio-genomics analysis of I-SPY2 trial

Background

The pan-Akt inhibitor MK-2206 has shown promising results in improving the pathological complete response (pCR) rate in high-risk early-stage breast cancer in the I-SPY2 trial. However, a considerable portion of patients does not respond to MK-2206 in addition to standard therapy. This study aimed to discover image biomarkers from DCE-MRI images that can further select responders of MK-2206.

Methods

A total of 1104 patients from the I-SPY2 trail were enrolled in this study. DCE-MRI images were collected from the 92 patients in the MK-2206 arm and 209 in the control arm for biomarker discovery. Image biomarkers, extracted from pre-treatment DCE-MR images within the functional tumor volume, were identified with high repeatable/reproducible and significant associations between pCR and biomarker-treatment interaction. Treatment sensitivity of image biomarker-defined subtypes were quantified with odds ratio (OR). Bayesian logistic regression was used to estimate the pCR rates of the MK-2206 and control arm. An exploration of differentially expressed proteins/genes and biological pathways was also conducted.

Results

One reliable image biomarker, glcm_SumSquares (GLCM_SS), was found to be predictive of treatment response. The positive group (GLCM_SS+) was highly sensitive to MK-2206 in the entire discovery cohort and specific subgroups (HER2-, HR-, HER2-/HR-, and MammaPrint (MP)2) with ORs ranging from 5.50 to 18.13 (P<0.001). With the image biomarker, the mean estimated pCR rates of the MK-2206 arm increased significantly from 38% to 56% (P=0.022) in the entire discovery cohort, 30% to 46% (P=0.023) in the HER2- subtype, 49% to 71% (P=0.005) in the HR- subtype, 39% to 63% (P=0.009) in the HER2-/HR- subtype, and 44% to 64% (P=0.005) in the MP2 subtype. GLCM_SS+ was also associated with overexpression of total PTEN protein and enriched by immune-related pathways.

Conclusions

An image biomarker, GLCM_SS, has been identified to be able to select the responders of AKT-inhibitor MK-2206. GLCM_SS+ also showed distinct gene profiles enriched in immune signaling. Further experiments are planned to verify the radiogenomics association with GLCM_SS+ sensitivity to MK-2206.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

Jing Cai.

Funding

Mainland-Hong Kong Joint Funding Scheme (MHKJFS) (MHP/005/20); Shenzhen Basic Research Program (JCYJ20210324130209023); Project of Strategic Importance Fund (P0035421) and Projects of RISA (P0043001) from The Hong Kong Polytechnic University; Health and Medical Research Fund (HMRF 09200576), the Health Bureau, The Government of the Hong Kong Special Administrative Region.

Disclosure

All authors have declared no conflicts of interest.