Imaging of bevacizumab treated brain: traditional and emerging concepts Asim K. Bag Joel K Cure Aparna Singhal David Wever Asim K. Bag Joel K Cure Aparna Singhal David Wever

AK Bag: No disclosure J Cure: No disclosure A Singhal: No disclosure D Wever: No disclosure

Outline Physiology of bevacizumab Indication of bevacizumab in glioblastoma multiforme (GBM) Imaging changes induced by bevacizumab Based on RANO criteria Persistent diffusion restriction

What is bevacizumab? Bevacizumab is a recombinant humanized monoclonal antibody against vascular endothelial growth factor (VEGF).

Mechanism of action Bevacizuamb directly binds to VEGF extracellularly to prevent interaction of VEGF with VEGF receptors (VEGFRs) on the surface of endothelial cells. The VEGFR family is primarily responsible for pro-angiogenic and other tumorigenic VEGF signaling. Bevacizuamb blocks the downstream effects of VEGF signaling.

Mechanism of action Blockade of normal VEGF action can lead to Regression of existing tumor vessels This leads to reduction of tumor size Inhibition of formation of new tumor vessels This leads to potential inhibition of tumor growth.

Indication of bevacizumab in GBM Based on demonstration of durable objective response rates observed in two single-arm trials, AVF3708g and NCI 06-C-0064E, the FDA granted accelerated approval to bevacizumab injection as a single agent for patients with glioblastoma, with progressive disease following prior therapy.

Imaging changes induced by bevacizumab in the setting of GBM Tumoral response based on RANO criteria: Complete response Partial response Stable tumor Progressive disease Pseudoresponse Tumoral and Extratumoral response Persistent diffusion restriction Progressing Improving Stable

GBM: Complete response Complete Response according to RANO criteria 1 : Defined by ALL of the following: 1.Complete disappearance of all enhancing measurable sustained for at least 4 weeks; 2.No new lesions; 3.Stable or improved non-enhancing (FLAIR) lesions; 4.Patient was on steroids with physiologic replacement doses only; 5.Patient improved clinically. 1 J Clin Oncol 28:

Pre-Avastin 2 mo post-Avastin5 mo post-Avastin Complete response This patient met all of the RANO criteria for complete response: 1.Complete disappearance of all enhancing measurable sustained for at least 4 weeks; 2.No new lesions; 3.Improved non-enhancing FLAIR lesions; 4.Was on physiologic replacement doses only; 5.Patient improved clinically. 1 J Clin Oncol 28:

GBM: Partial response Partial Response according to RANO criteria 1 : Defined by ALL of the following: 1.Fifty percent decrease compared with baseline in the sum of products of perpendicular diameters of all measurable enhancing lesions sustained for at least 4 weeks; 2.No progression of non-measurable disease; 3.No new lesions; 4.Stable or improved non-enhancing FLAIR lesions on same or lower dose of corticosteroids compared with baseline scan (corticosteroid dose at the time of the scan evaluation should be no greater than the dose at time of baseline scan); 5.Stable or improved clinically. 1 J Clin Oncol 28:

Pre-Avastin 2 mo post-Avastin5 mo post-Avastin Partial tumor response This patient met all RANO criteria for partial tumor response: 1.50% decrease in sum of the products of perpendicular diameters of enhancing lesions compared to baseline 2.NO new lesion 3.Improved non-enhancing FLAIR lesion 4.Stable clinically

GBM: Progression of tumor Progression of tumor according to RANO criteria 1 : Defined by ANY of the following: 1. 25% increase in sum of the products of perpendicular diameters of enhancing lesions compared with the smallest tumor measurement obtained at baseline 2. Significant increase in FLAIR non-enhancing lesion on increasing doses of corticosteroids compared with baseline scan 3. New lesion 4. Clinical deterioration not attributable to other causes apart from the tumor 1 J Clin Oncol 28:

Progression of tumor This patient met 3 of the 4 RANO criteria for progression of tumor: 1.25% increase in sum of the products of perpendicular diameters of enhancing lesions compared to baseline 2.Significant increase in non-enhancing FLAIR lesion 3.Clinical deterioration not attributable to other causes apart from the tumor Pre-Avastin 2 mo post-Avastin4mo post-Avastin

GBM: Stable tumor Stable tumor according to RANO criteria 1 : Defined by ALL of the following: 1. Does not qualify for complete response, partial response, or progression; 2.Stable non-enhancing FLAIR lesions on same or lower dose of corticosteroids compared with baseline scan. 1 J Clin Oncol 28:

Pre-Avastin 2 mo post-Avastin4 mo post-Avastin Stable tumor This patient met RANO criteria for stable tumor: 1.No change in enhancement 2.No new lesion 3.Stable –to-slightly improved non- enhancing FLAIR lesion 4.Stable clinically

GBM: Pseudoresponse Bevacizumab may produces a rapid decrease in the degree of enhancement, sometimes within hours of beginning therapy 1. Can demonstrate radiologic response in up to 50% of GBMs based on McDonald criteria 2. These apparent responses may be partly a result of normalization of abnormally permeable tumor vessels and not always necessarily indicative of a true antiglioma effect. RANO 1 criteria suggests that radiologic responses should persist for at least 4 weeks before they are considered as true response. 1J Clin Oncol 28: J Clin Oncol 8:

Pre-Avastin 3 weeks post-Avastin 8 weeks post-Avastin Pseudoresponse This patient met RANO criteria for pseudoresponse of tumor: 1.Significantly improved enhancement, non- enhancing FLAIR abnormality at 3 weeks 2.Worsening mass effect and FLAIR abnormality at 8 weeks. Note: The degree of enhancement at 8th week is less intense but is more patchy and diffuse compared to the baseline scan suggesting change in tumor morphology.

Persistent diffusion restriction Persistent diffusion restriction frequently develops after initiation of bevacizumab treatment. The exact mechanism is still not known. Persistent diffusion restriction could be due to tumor progression or due to atypical necrosis due to chronic hypoxia induced by blockade of tumoral/peritumoral blood vessels or due to combination of both of these. Development of the persistent diffusion restriction has been associated with better survival.

Persistent diffusion restriction contd…. Biopsy of the area with persistent diffusion restriction has shown Gelatinous necrotic tissue 1 Prominent fibrosis of the blood vessels 1 Upregulation of HIF-1alpha in endothelial, neuroepithelial and inflammatory nuclei 2 Hyalinized and fibrotic blood vessels 2 Presence of sideroblast, cholesterol clefts and multinucleated giant cells

Persistent diffusion restriction contd…. Persistent diffusion restriction is frequently associated with T1 shortening Susceptibility effects on SWI and GRE sequences This effect could be due to presence of iron in some form in the tissue (sideroblast) Volume of area of persistent diffusion restriction has been shown to be stable up to 90% of cases

Persistent diffusion retsriction In addition to stable area of diffusion restriction, we are showing dynamic nature of the area of the diffusion restriction Enlargement of area of diffusion restriction. Reduction of area of diffusion restriction.

Enlargement of Atypical necrosis Enlargement of area of diffusion restriction. We believe that this is secondary to enlargement of the necrotic area because Progressive enlargement of the area of T1 shortening within the core of diffusion restriction with clearing of more central areas (see next slide) Absence of worsening of mass effects over time in spite of enlargement of the area of diffusion restriction The patient is stable clinically 2 years after the initiation of bevacizumab therapy No increased rCBV in the area of persistent diffusion restriction.

Pre-Avastin 2 mo post-Avastin 6 mo post-Avastin 8 mo post-Avastin11 mo post-Avastin 21 mo post-Avastin Enlargement of atypical necrosis. The diffusion restriction is not associated with increased perfusion. The patient is alive 23 months after initiation of Avastin therapy.

Decreased size of Atypical necrosis Reduction of area of diffusion restriction. We believe that this is secondary to change in the morphology of the atypical necrosis, likely due to more gliotic component. This is supported as because Progressive enlargement of cystic encephalomalacia as seen on T1 and FLAIR images. Absence of worsening of mass effects over time in spite of enlargement of the area of diffusion restriction. No increased rCBV in the area of persistent diffusion restriction.

Pre-Avastin 2 mo post-Avastin 18 mo post-Avastin 30 mo post-Avastin Metamorphosis of the atypical necrosis with more gliotic component Reversal of diffusion restriction Progressive enlargement of cystic necrosis as evidenced on T1 and FLAIR images Very low rCBV Absence of mass effect

Stable size of Atypical necrosis

Pre-Avastin 7 mo post-Avastin 11 mo post-Avastin 17 mo post-Avastin Stable Atypical necrosis

Conclusions Treatment of with bevacizumab causes different types of imaging appearance. Changes can be Classical Comple response, Partial response Tumor progression Stable tumor Pseudoresponse Atypical necrosis Enlarging Decreasing Stable