Radiation and Cancer Physics
PD 13 - Physics 8 - Poster Discussion - Outcome Analysis and Response Imaging
1116 - Comparison of Radiographic Approaches to Assess Treatment Response in Pituitary Adenomas: Are RECIST/Rano Good Enough?
Tuesday, October 23
5:33 PM - 5:39 PM
Location: Room 217 C/D
Comparison of Radiographic Approaches to Assess Treatment Response in Pituitary Adenomas: Are RECIST/Rano Good Enough?
B. S. Imber1, T. J. Yang2, Z. Zhigang2, K. Beal2, M. Cohen3, V. S. Tabar4, E. B. Geer2, A. Lin2, and R. J. Young2; 1Department of Radiation Oncology, Memorial Sloan Kettering Cancer Center, New York, NY, 2Memorial Sloan Kettering Cancer Center, New York, NY, 3Department of Surgery, Memorial Sloan Kettering Cancer Center, New York, NY, 4Multidisciplinary Skull Base and Pituitary Center at Memorial Sloan Kettering Cancer Center, New York, NY
Purpose/Objective(s): A gold standard for radiographic post-treatment response (PTR) of pituitary adenomas (PA) does not exist; response criteria are needed for management decisions and clinical trial design. RECIST assesses PTR based on longest diameter. RANO, the criteria used for intracranial tumors, measures size by the product of perpendicular diameters. In this study, we evaluate these surrogate approaches, in addition to calculated volumes.
Materials/Methods: This study evaluates 35 consecutively treated macroadenoma patients who received pituitary radiotherapy (RT). A single experienced neuro-radiologist estimated tumor sizes on the pre-RT MRI and first post-RT MRI using 4 surrogate methods: 1D per RECIST criteria, 2D per RANO criteria, estimated 3D using measured diameters assuming a perfect sphere (3DS) and estimated 3D assuming a perfect ellipsoid (3DE). Each tumor was manually segmented (TeraRecon software) for a 3D volumetric (3DV) size and the Hakon-Wadell method was used to assess the shape’s sphericity. PTR was defined as change in post-RT tumor size vs. pre-RT. Each surrogate PTR was statistically compared against the assumed 3DV gold standard using Pearson’s correlation and intraclass correlation coefficient (ICC). PTR was dichotomized as either “progression” or “non-progression” using accepted criteria to calculate sensitivity/specificity.
Results: All patients (median age = 51 years; 51% male) completed RT (median/mean 50 Gy, range 45-54) for pathologically confirmed PA. Patients had a median of 1 (range 0-4) resection a median of 6.1 months before RT. Median duration between scans was 5 months. Using 3DV segmentation, median size of the pre-RT tumors was 6.8 cm3 (range 0.5-51.6, std dev 10.4). Median sphericity of the pre-RT volumetric tumors was 0.63 (range 0.39-0.79, std dev 0.11). Median PTR by 3DV was -9% and 7/35 patients (20%) had a >20% increase in measured volume, whereas 12/35 (34%) had a >20% decrease in volume. The PTR estimated by the 1D, 2D and 3DS methods were overall similarly correlated to the assumed 3DV gold standard (e.g., 1D Pearson correlation 0.67, p<0.0001) and were all superior to the 3DE approach. The ICC demonstrated moderate to good reliability of the PTR estimated by the 1D (ICC =0.54, p <0.001), 2D (ICC = 0.60, p<0.001) and 3DS (ICC = 0.53, p <0.001) methods while the 3DE method was slightly inferior (ICC = 0.39, p=0.009). Using 3DV, based on the +20% PTR progression threshold (per literature review), 7 patients progressed on post-RT MRI. Using 1D with a +20% PTR threshold (per RECIST), compared to 3DV, we estimate 29% sensitivity and 96% specificity. Using 2D with a +25% PTR threshold (per RANO), compared to 3DV, we estimate a 14% sensitivity and 89% specificity.
Conclusion: Assessing response based on 2D and calculated 3D volume did not improve upon one dimensional assessment. Treated PA have low sphericity, resulting in low sensitivity at capturing a 20% change in volume. A response criteria for making management decisions should be based on 3D measured volumes.
Author Disclosure: B.S. Imber: None. T. Yang: None. K. Beal: Stock; Magnolia Medical Technology (MMT). M. Cohen: None. E.B. Geer: None. A. Lin: None.