Digital quantification of Ki-67 and PHH3 in the classification of uterine smooth muscle tumors

Abstract

<h3>Objectives:</h3> The aims of this study were to evaluate if digital quantification of proliferation marker Ki-67 and mitotic marker phosphohistone H3 (PHH3) can reliably distinguish leiomyosarcoma (LMS) from smooth muscle tumors of uncertain malignant potential (STUMP), leiomyoma with bizarre nuclei (LMBN), and leiomyoma (LM), and if digital quantification of these markers correlates with clinical outcomes. <h3>Methods:</h3> Cases of LMS, STUMP, LMBN, and LM for which paraffin-embedded tissue was available were identified from a single institution. Manual evaluation of mitotic count was performed using hematoxylin and eosin (H&E) evaluation on 50 high powered fields (5 mm²). Immunohistochemical staining for Ki-67 and PHH3 was performed on one section from each case, and a high throughput digital quantification system was used to calculate an average percent stain. Receiver operating characteristic (ROC) curves for the ability of Ki-67 and PHH3 to predict LMS versus the other diagnostic groups were generated. Univariate Cox regression assessed for associations between each marker and progression-free survival (PFS) among patients with available follow-up. <h3>Results:</h3> A total of 39 cases (17 LMS, 5 STUMP, 10 LMBN, and 7 LM) were included. Mitotic count, Ki-67, and PHH3 were significantly correlated. When comparing the LMS group to the STUMP, LMBN, and LM groups combined, LMS showed a significantly greater digital quantification of Ki-67 (median 10.6% versus 0.4%, p<0.001), and PHH3 (median 0.5% versus 0.14%, p=0.022). Area under the receiver operating characteristic curve (AUC) tests suggested Ki-67 was a better predictor of LMS than PHH3 (AUC 0.92 vs 0.73, p=0.017) (Figure 1). Above a threshold Ki-67 value of 3.8%, the sensitivity was 82% and specificity was 91%. Among 10 LMS/STUMP patients with available follow-up (8 LMS, 2 STUMP), inferior progression-free survival was noted for patients with higher values of Ki-67 (hazard ratio [HR] 1.099, p=0.039) and PHH3 (HR 2.43, p=0.035). <h3>Conclusions:</h3> This study provides preliminary evidence that digital quantification of Ki-67, and possibly PHH3, can aid in a fast and reliable diagnosis of LMS versus other uterine smooth muscle tumors.