Subjective Pathologic Estimates of Viable Tumor in Ablated Hepatocellular Carcinomas (HCC) Are Adequate for Routine Practice and for Radiology/Pathology Correlation Studies
Joshua A Hanson, Alison VanDyke, Joanna Gibson, Marie Robert, Dhanpat Jain, Kisha A Mitchell. Yale New Haven Hospital, New Haven, CT
Background: Ablation of HCC is employed to induce complete tumor necrosis. Incomplete necrosis predicts HCC recurrence and may affect transplant options. Clinicians rely on radiologists and pathologists to accurately asses treatment response. Studies examining radiology/pathology correlation in determining complete response have shown poor agreement, however, the pathologic methods used to assess viable tumor (VT) are often not uniform or are poorly described. We compared subjective and objective measurements of the percentage of viable tumor (PVT) in ablated HCCs to determine if subjective estimates are reliable.
Design: Two pathologists independently reviewed gross images and H&E stained slides for 23 ablated HCCs. The livers were serially sectioned and ablation cavities (AC) were measured in three dimensions. VT was defined as non-necrotic tumor within or contiguous with an AC. Each observer subjectively estimated PVT after gross and microscopic review (eyeball estimate). Each observer then measured individual areas of VT foci using formulas based on their shapes and added them to yield a composite viable tumor area (CVTA) for each AC. The CVTA was converted to a composite viable tumor volume (CVTV) based on the spherical nature of the AC. The CVTV was divided by the volume of the ablation cavity as determined by gross measurements, yielding an objective estimate of PVT. Means and standard deviations were calculated for all estimates and paired sample t-tests were conducted examining the inter and intraobserver agreements.
Results: The 2 pathologists had strong interobserver agreement of PVT using both the subjective [means: 25.7%(KM), 25.3%(JH); p=0.81] and objective methods [means 21.5%(KM), 21.6%(JH); p=0.75]. The intraobserver agreement between subjective and objective methods was less strong [p=0.09 (Km), p=0.05 (JH)] without a clear trend. Data analysis revealed 5 outliers with poor intraobserver agreement. On review, less than 1 section per cm of AC was submitted for each of these 5 cases. The paired sample t-tests were re-calculated without the poorly sampled ACs and yielded stronger intraobserver agreement between the two methods [p=0.38(KM), p=0.69(JH)].
Conclusions: This study demonstrates that a subjective assessment of PVT correlates well with measured PVT. It therefore validates the subjective pathologic assessment of HCC ACs as an accurate measure of the PVT. It also emphasizes the importance of adequate sampling of the AC to ensure accurate estimates.
Tuesday, March 20, 2012 9:30 AM
Poster Session III # 253, Tuesday Morning