Cerebral blood volume and apparent diffusion coefficient – Valuable predictors of non-response to bevacizumab treatment in patients with recurrent glioblastoma

Abstract

Background: Glioblastoma multiforme (GBM) is the most common primary brain tumor in adults. The core of standard of care for newly diagnosed GBM was established in 2005 and includes maximum feasible surgical resection followed by radiation and temozolomide, with subsequent temozolomide with or without tumortreating fields. Unfortunately, nearly all patients experience a recurrence. Bevacizumab (BV) is a commonly used second-line agent for such recurrences, but it has not been shown to impact overall survival, and short-term response is variable. Methods: We collected MRI perfusion and diffusion images from 54 subjects with recurrent GBM treated only with radiation and temozolomide. They were subsequently treated with BV. Using machine learning, we created a model to predict short term response (6 months) and overall survival. We set time thresholds to maximize the separation of responders/survivors versus non-responders/short survivors. Results: We were able to segregate 21 (68%) of 31 subjects into unlikely to respond categories based on Progression Free Survival at 6 months (PFS6) criteria. Twenty-two (69%) of 32 subjects could similarly be identified as unlikely to survive long using the machine learning algorithm. Conclusion: With the use of machine learning techniques to evaluate imaging features derived from pre-and post-treatment multimodal MRI, it is possible to identify an important fraction of patients who are either highly unlikely to respond, or highly likely to respond. This can be helpful is selecting patients that either should or should not be treated with BV.

Predicting Survival in Glioblastoma Patients Using Diffusion MR Imaging Metrics—A Systematic Review

“…Patel et al [ 40 ] also investigated the utility of ADCL, concluding that patients with low ADCL have a survival benefit when surgically excised, whereas large tumours with high ADCL may be best treated with BV. Three studies investigated the survival predictive power of the ADC monomodal histogram parameters [ 39 , 58 , 78 ], of which two involved BV-treated patients [ 58 , 78 ]. In the first it was found that, unlike the normalized ADC 5th percentile that was also evaluated, the volume of low-ADC lesions in enhancing pre- and post-BV scans predicted shorter OS [ 58 ], while in the second [ 78 ], using a machine learning approach, the ADC histogram metrics in combination with perfusion metrics arising from pre- and post-treatment images were found to be useful for BV response assessment in terms of PFS and OS.…”

“…Three studies investigated the survival predictive power of the ADC monomodal histogram parameters [ 39 , 58 , 78 ], of which two involved BV-treated patients [ 58 , 78 ]. In the first it was found that, unlike the normalized ADC 5th percentile that was also evaluated, the volume of low-ADC lesions in enhancing pre- and post-BV scans predicted shorter OS [ 58 ], while in the second [ 78 ], using a machine learning approach, the ADC histogram metrics in combination with perfusion metrics arising from pre- and post-treatment images were found to be useful for BV response assessment in terms of PFS and OS. Among the histogram features inspected by Zolal et al [ 39 ] in R-GBM patients undergoing their second surgery, the ADC histogram skewness was an independent prognostic factor for OS and PFS after the second surgery.…”

“…Continued research efforts and standardization of acquisition parameters and analytic methods, possibly with automation, are required to arrive at the most effective approach that can be applied across institutions. Moreover, due to the ever-growing application of artificial intelligence techniques in the context of medical imaging, it could be interesting to further explore the association between diffusion metrics and GBM survival outcomes using machine learning and deep learning techniques, as done by Chang et al and Petrova et al for DWI metrics [ 65 , 78 ] or by Zacharaki et al and Nie et al [ 92 , 93 ] for DTI metrics, even if the latter two works were not specific for GBM patients.…”

Neuroimaging Studies Evaluating Effective Therapies for High-Grade Glioma

“…This approach could be very helpful for patient stratification into subgroups that are likely/unlikely to respond to Bevacizumab therapy and, thus, for therapeutic planning. 187 Another study by Stecco et al 188 demonstrated the reliability of rCBV as predictive biomarker of PFS in recurrent GBM patients.…”

“…In the last couple of years there has been growing research evidence showing that the large EPR effect observed in preclinical mouse models cannot be reproduced in the clinic. 187 published a study involving the use of machine learning techniques to evaluate changes in MRI parameters of apparent diffusion coefficient (ADC) and relative CBV (rCBV) associated with the likelihood of response to Bevacizumab therapy in recurrent GBM patients. This approach could be very helpful for patient stratification into subgroups that are likely/unlikely to respond to Bevacizumab therapy and, thus, for therapeutic planning.…”