Comparison of Semi-Quantitative Versus Quantitative Grading System in Endocrine Tumors of the Pancreas: Which One Should Be Applied?
Birgit Weynand, Ivan Borbath, Christine Sempoux, Jean-Francois Gigot, Anne Jouret-Mourin. Cliniques Universitaires St Luc, Brussels, Belgium
Background: Pancreatic endocrine tumors (PETs) are rare, representing 1-2% of all pancreatic tumors. ENETS and WHO have defined criteria for a grading system for PETs to assess their prognosis based on mitotic count or Ki67 labeling index (Ki67-LI). Mitotic count should be done on at least 40 fields at 40x magnification (10 HPF=2 mm2) in areas of highest mitotic density. Ki67-LI should be assessed in 500-2,000 tumor cells in areas of highest nuclear labeling. This recommendation is difficult to implement in routine practice, because of the time-consuming aspect. Therefore many pathologists use a semi-quantitative approach for grading.
Aim: To compare semi-quantitative to quantitative assessment of PETs 'grading system.
Design: 36 surgically resected PETs from 35 patients were assessed after Ki67 immunohistochemistry. All were counted by 2 independent pathologists (BW=observer 1 and AJM=observer 2) (mean number of counted cells: 2001±49; minimum: 1480, maximum: 2130). The semi-quantitative assessment was done by scanning the tumor at low and high power. The same pathologists did this second evaluation after a 6 months interval.
Results: Semi-quantitative assessment was not concordant for 9 of 36 tumors (25%). Eight cases were discordant between G1 and G2 and one between G2 and G3. In 7 cases, observer 2 was higher than observer 1 and in 2 cases it was the opposite. When assessing Ki67-LI, discordance was observed in 1 case out of 36 PETs (2.8%), in which observer 1 classified the tumor as G1, whereas for observer 2, it was a G2. Interestingly, this case was discordant in both grading systems. Inter-observer correlation was very good and statistically highly significant for Ki67-LI (Pearson r2= 0.95, p<0.001). Whereas, it was lower in the semi-quantitative assessment (Pearson r2= 0.57, p<0.001)
Conclusions: Grading system for PETs is difficult to apply in routine practice because it is time-consuming. Therefore, many pathologists tend to use a semi-quantitative approach of mitotic count or Ki67 positivity in PETs. In this study on 36 PETs, it is shown that inter-observer correlation is much better when counting Ki67-LI. We therefore advocate all pathologists to count effectively Ki67 positive cells to assess prognosis of PETs.
Wednesday, March 21, 2012 9:30 AM
Poster Session V # 252, Wednesday Morning