Herein, we conducted a retrospective study based on our previous phase 3 trial and revealed that high Ki67 expression before treatment could predict survival benefits from the addition of TPF neoadjuvant chemotherapy in patients with locally advanced OSCC.
The prognostic value of Ki67 has been widely investigated in various kinds of cancers while they exhibited conclusions different from one another [8, 11, 15]. Ki67 detected in the surgical specimens was usually defined as a negative prognostic factor in some studies [8, 16], where surgery alone or plus post-operative therapy based on pathological assessment was performed. This may be explained that the residual cancer cells with high proliferative ability after surgery are more likely to cause recurrence.
On the other hand, some studies considered Ki67 as a positive prognostic factor. The common features of these studies were the introduction of radiotherapy or chemotherapy into the treatment regimen and the detection of Ki67 before treatment [11, 12, 17, 18]. As tumors with higher Ki67 expression were proved to be more sensitive to radiotherapy and chemotherapy, it could be easily understood that they would have better response and prolonged survival upon the specific treatment. It was retrospectively reported that patients with high basal Ki67 expression in breast cancer were prone to have complete pathological response (pCR) after neoadjuvant therapy, and it could act as a positive prognosis factor in pCR patients [10, 14]. The decline of Ki67 expression after neoadjuvant therapy could also reflect the treatment effect and had even more favorable predictive value of prognosis [9, 18–20]. Besides the two opinions above, there are still a proportion of studies that failed to prove any prognosis value of Ki67 [21, 22].
Although contradictory conclusions were drawn from the previous studies, it could become explainable when considering the heterogeneous treatment regimens of the study population. It has been confirmed that patients with high Ki67 expression before treatment would obtain survival benefits from chemotherapy and radiotherapy. However, the correlation between Ki67 and prognosis would be led to the opposite direction when the study population were mixed up with those who received surgery alone, where Ki67 represented the potential to relapse. Therefore, the conclusions would vary among positive, negative and no significant value depending on the mixing ratio of different treatment strategies. Besides, other factors may also contribute to the controversy, such as the different Ki67 cut-off values and different subtypes of cancers [23].
In a previous retrospective study, Klimowicz et al [12] enrolled OSCC of all stages and reported Ki67 as a marker for good prognosis in patients with OSCC, and the prognostic impact was more pronounced in the postoperative radiotherapy-treated subgroup. As most of the patients (n = 83, 69%) received surgery plus post-operative radiotherapy, it was reasonable to speculate the decisive effect of the proportion of radiotherapy-treated population and further verified the explanation above. Therefore, it is vital to ensure consistency of treatment regimen when assessing the predictive value of Ki67.
To our knowledge, this is the first study to explore the prognostic value of Ki67 in patients with locally advanced OSCC treated with TPF neoadjuvant therapy prior to standard therapy. We enrolled the patients from the experimental arm of our previous phase 3 trial, in which patients received the same treatment of TPF neoadjuvant therapy, surgery and post-operative radiotherapy. The consistency of the treatment regimen could enhance the credibility of the conclusion.
However, there are still some limitations of the present study. Firstly, we lacked the investigation of pathological response and Ki67 expression after chemotherapy. Even though patients with high level of Ki67 were proved to have better clinical outcomes as a group, there may still exist a proportion of patients that possessed resistance to the TPF neoadjuvant therapy. Pathological response and Ki67 decline after medication could reflect the treatment effect and identify the insensitive population, which we failed to demonstrate in this study. Secondly, Ki67 may have a stronger predictive value when combined with other biomarkers than itself alone. It was reported that combining Ki67 with residual cancer burden following neoadjuvant chemotherapy in breast cancer provided significantly more prognostic information than either alone [24]. Thus, Ki67 could be a candidate for a predictive panel of OSCC and required further exploration.