Isolation of low grade urinary bladder tumors carrying bad prognosis using KI-67 labeling index

Introduction: Even among low grade bladder tumors, there are some which have been found to carry bad prognosis for progression, recurrence and survival. Accurately pinpointing these cases at the time of diagnosis just with the help of routine histopathology is problematic. Ki-67 immunostaining may potentially be of help in identifying such cases, making administration of additional treatment modalities like chemotherapy and radiotherapy possible. There is scarcity of such studies in our country. The present study aims to correlate Ki-67 labeling index with the histopathological type and grade, pathological stage and mitotic index of urinary bladder tumors. Objective: Objective of the study is to consider the possibility of using Ki-67 labeling index to identify those low grade urinary bladder tumors which carry bad prognosis. Materials and Methods: All trans-uretheral resection of bladder tumor (TURBT) specimens in our institute were analyzed for their histopathological grade, pathological stage, mitotic index and Ki-67 labeling index in the period extending from from August 2015 to December 2016. The clinical data were analyzed retrospectively and statistically. SPSS Version 23 (SPSS Inc.,Chicago, IL, USA) software was utilized for all statistical analyses. Chi-square analysis was utilized for comparison of groups. Results with P <0.05 were taken as statistically significant. Results: Ki-67 expression in Low grade urothelial carcinoma (LGUC) showed variation from 1.4% to 41%. The papillary urothelial neoplasm of low malignant potential (PUNLMP) cases showed variation from 16 -17%. And the only case of transitional cell papilloma in the study was found to have expression of 9%. Conclusion: Cases of high (more than 20%) Ki-67 MIB-1 labeling index can be segregated even from among those cases with low histopathological grade, low mitotic index and pTa histopathological stage. Such cases are candidates for multimodality treatment including radiotherapy, chemotherapy etc. © This is an open access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/) which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.


Introduction
According to statistics from the Indian Cancer Registry, bladder cancer is the ninth most common cancer recorded in India, contributing 3.9% of all cancer cases, regardless of gender. 1 The histopathological grade of bladder tumors is usually critical in determining the clinical management and prognosis of patients. However, it has been seen that even among low grade bladder tumors, there are some which carry bad prognosis for progression, recurrence and survival. Segregating these cases with the help of routine histopathology is problematic. However, identification of such cases can be made easier with the help of Ki-67 immunostaining. Besides, Ki-67 labeling index also proves helpful in assessment of urothelial tumors where a controversy exists in labeling a tumor low grade or high grade by other methods. 2 https://doi.org/10.18231/j.pjms.2020.041 2249-8176/© 2020 Innovative Publication, All rights reserved. 187 There is a dearth of literature in our country on the utility of Ki-67 immunostaining in urinary bladder tumors. The present study has tried to fulfill this lacuna.

Materials and Methods
The present study is a retrospective institutional based study of paraffin embedded tissue of 80 cases of bladder tumors that came for treatment in our institution from August 2015 to December 2016. The specimens were received in the form of trans-urethral resection of bladder tumor (TURBT) chips. Specimens were fixed in 10% buffered neutral formalin for a period of 24 hours and subsequently the entire specimen was submitted for processing.
Four micron thick sections were made from paraffin blocks and:-1. Stained with Haematoxylin and Eosin for the assessment of histopathological type and grade, pathological stage, and mitotic index (calculated as number of mitotic figures per 10 high power fields). 2. Immunohistochemical analysis done using MIB-1 antibody provided by Biogenex©, US; following the instructions as provided in the kit.
Ki-67 immunoquantitation was performed utilizing light microscopy at x1000 magnification. Ki-67 labeling index (percentage of tumor cells displaying nuclear staining) was ascertained by counting at least 1000 tumor cells in 10 selected fields displaying the highest immunoreactivity. The histopathological grade, pathological stage and mitotic index were subsequently correlated with K-67 labeling index, cut value for which was taken as >20%. SPSS Version 23 (SPSS Inc., Chicago, IL, USA) software was utilized for all statistical analyses. Chi-square analysis was used for comparison of groups. Results with P <0.05 were accepted as statistically significant.
Out of 32 cases of high grade urothelial carcinoma, 24 (75%) cases were found to be of High grade papillary urothelial carcinoma, 6 cases (19%) were of high grade urothelial carcinoma with squamous differentiation, while 1 case (3%) was of high grade urothelial carcinoma with glandular differentiation and 1 case (3%) of High grade nonpapillary urothelial carcinoma (Nested variant).
With the exception of 1 case of transitional cell papilloma and 2 cases of papillary urothelial neoplasm of low malignant potential, invasion was studied in the remaining 77 cases. Out of 77 cases, 22 cases (28.6%) were of low grade papillary urothelial carcinoma in which no invasion was observed, 10 cases (13%) were of Low grade papillary urothelial carcinoma with only lamina propria invasion. Five cases (6.5%) were of Low grade papillary urothelial carcinoma which demonstrated muscularis propria invasion. Nine cases (11.7%) were of High grade papillary urothelial carcinoma in which no invasion was seen, two cases (2.6%) which demonstrated only lamina propria invasion, while 21 cases (27.2%) were of High grade papillary urothelial carcinoma which demonstrated muscularis propria invasion.
This study had no cases in pT3 and pT4 category as all cases were sourced through trans-urethral resection.
As can be seen in Table 1, the value of Pearson correlation coefficient was observed to be +0.725 which indicates strong positive correlation between histopathological grade and Ki-67 labeling index and the P value = 0.000, which demonstrates that the results are statistically significant. Figures 1 and 2 show examples of low and high Ki-67 nuclear positivity respectively.  Also as demonstrated in Table 1, six of the total 41 low grade bladder tumor cases in this study had Ki-67 labeling index above 20%. On the other hand, five out of 39 high grade cases had a Ki-67 labeling index below 20%.
As can be seen in Table 2, the value of Pearson correlation coefficient was observed to be =0.423 which indicates moderate correlation between pathological stage and Ki-67 labeling index, and the P value = 0.000, which demonstrates that the results are statistically significant.
Also, among the non-invasive (pTa) tumors, 10 out of total 34 such cases had above 20% Ki-67 labeling index. It was found that, even though most of such tumors were of high grade, two low grade tumors with pTa stage also had a high Ki-67 labeling index. Similarly three low grade tumors with pT1 stage and one low grade tumor with pT2 stage, also had a high Ki-67 labeling index.
On the contrary, as many as 8 out of total 33 cases in pT2 stage were found to have a Ki-67 labeling index of less than 20%. Thus, the correlation of Ki-67 index with pathological stage is less than its correlation with histopathological grade (as noted IN Table 2), and with mitotic index (as discussed in Table 3).
As can be seen in Table 3, the value of Pearson correlation coefficient was observed to be +0.759, which indicates strong positive correlation between mitotic index and Ki-67 labeling index, and the P value = 0.000, which demonstrates that the results are statistically significant. Table 3 also demonstrates that none of the cases with mitotic index of more than 20/10 hpf had a Ki-67 labeling index below 20%. But there were occasional (n=3) cases with low mitotic index (<10/10 hpf), that had a Ki-67 labeling index above 20%.

Discussion
Ki-67 is a nuclear protein which is present during G1, S, G2 and M phases of continuously cycling cells, though it is not present in G0 cells. Though not well characterized, the genetic locus of Ki-67 has been assigned to chromosome 10. It has been shown by several studies that cell proliferative activity as defined by Ki-67 labeling index correlates with the tumor growth fraction. MIB-1 monoclonal antibody detects a formalin-resistant epitope of Ki-67 antigen thereby making it suitable for use on routinely processed paraffinembedded tissue. 3 Different Ki-67 equivalent antibodies recognize different epitopes of Ki-67 antigen and, therefore, have dissimilar sensitivity and specificity. While comparing different available antibodies, Lindboe et al observed that MIB-1 antibody had a higher sensitivity than other antibodies, and it also gave the best visual staining, with more diffusely and strongly stained nuclei. 4 The prototypic Ki-67 antibody had the limitation of being usable only on frozen section specimens. Moreover, it was observed that staining with MIB-1 led to demonstration of considerably higher labeling indices as compared to the prototypic Ki-67 antibody, probably due to better preservation of epitopes after formaldehyde fixation and antigen retrieval than by freezing the tissue. 5,6 The labeling index can be defined as the percentage of immunoreactive nuclei. 4 MIB-1 Ki-67 equivalent antibody was used in present study too to determine the labeling index.
In present study, Ki-67 MIB-1 labeling index was observed to be more than 20% in 14.6% of all the low grade  Other researchers have either not mentioned the cut off value in their studies or used a different cut off value. Santos et al, using more than 18% immunoreactivity as their cut off value observed that, in Grade 1, 29.2% patients came in this group and in Grade 2, 70.8% patients had more than 18% immunoreactivity. 9 Ding W et al, using Ki-67 immunoreactivity equal to or more than 25% as their criteria of Ki-67 positivity observed that 32.5% of cases with nonmuscle invasive bladder cancer were Ki-67 positive. 10 Tsuji M et al observed in their study that MIB-1 labeling index averaged 17.8 in Grade 2 and 32.3 in Grade 3 patients. 11 Goyal S et al observed Ki-67 index 11.8 ± 8.23% in low grade and 35.86 ± 17.55% in high grade papillary urothelial bladder cancer. These researchers also found that cut-off values for Ki-67 and mitotic indices can confirm high grade bladder carcinoma in equivocal cases. They also observed that Ki-67 and mitotic count can serve as potential and reliable indicators of muscle invasion. Thus they reported that Ki-67 equal to or more than 32.5% and 14 or more mitoses per 10 high power fields was 100% specific for high grade bladder carcinoma. Besides, Ki-67 equal to or more than 59% and 37 or more mitoses per 10 high power fields were 100% specific for muscle invasive bladder carcinoma. 12 In present study, also, Ki-67 MIB-1 labeling index values correlated well with mitotic indices and pathological staging of the tumor.
The Ki-67 labeling index value for low and high grade tumors by various researchers has been shown in tabular form in Table 4.
In another table, Table 5, comparison of Ki-67 labeling index values of present study with those of Margulis V et al 7 has been done according to age group, gender, tumor stage and tumor grade. While the age and sex distribution in both the studies is comparable, present study had comparatively higher percentage of patients in positive cohort in pT2 stage. Present study also had a higher percentage of Ki-67 labeling index positivity among the patients with high grade tumors.
In Table 6, a comparison of range of Ki-67 labeling index values for low grade and high grade tumors of present study with two other recent studies is shown.
Using Ki-67 expression, Gonul IT et al could distinguish "papillary urothelial neoplasm of low malignant potential" from other grade 1 urothelial tumors, thus obtaining more precise prognostic information. 14 There were two such cases in present study.
Only 33 out of 80 cases were non-invasive in present study whereas approximately 80% of bladder tumors have been found to be non-invasive in literature. 9,15 The underlying reason maybe that our patients present late to the hospitals for treatment. Any difference in biological behavior needs to be investigated.
Most of the patients in present study were males. This is in accordance with other reports in literature. 7,15 For superficial bladder tumors, classical histopathological markers do not define risk profiles and, consequently, there are no optimal treatment approaches and follow-up schedules. The course of disease of such tumors is, therefore, uncertain. 16 By adding Ki-67 index to the risk model, it became possible to ascertain subset of   low grade urothelial superficial papillary carcinomas with a worse course of disease, which can be closely followed up and may benefit from adjuvant treatments. 9 Three out of total 41 low grade tumors in present study fall in this group.
Radical cystectomy is the standard treatment for patients who have refractory non-muscle invasive and muscle invasive urothelial carcinoma of the urinary bladder. 17,18 Standard prognostic features like pathological stage and grade are of limited use in prediction of outcomes in these patients. Molecular biomarkers like Ki-67 may play a role in stratifying the heterogeneous patient population undergoing radical cystectomy into risk categories which can help in determining as to whether patients should be treated by observation or with adjuvant therapy. 19 Recently, Tanabe K et al have commented that muscle invasive bladder cancer patients with high Ki-67 expression status might derive benefit from chemoradiotherapy based multimodal approaches in terms of prognosis and quality of life as a result of bladder preservation. 20 Even though follow-up of the patients was not a part of present study, several studies have demonstrated an association between proliferation and tumor grade, stage, recurrence and prognosis. 21, 22 Quintero A et al concluded that tumor proliferation measured by Ki-67 labeling index is related to tumor recurrence, stage progression, and is an independent predictor of disease free survival, progression free survival and cancer-specific overall-survival in TaT1 bladder urothelial cell carcinoma. 15 Interestingly, a recent systematic review and metaanalysis has stated that significant correlations between high Ki-67 expression and survival outcome (recurrencefree/ progression-free/ overall cancer-specific survival) are present only in European -American patients and it does not reach significant level in Asian population. 23 Thus, further research in this regard in Asian countries is called for.
Some researchers have linked high Ki-67 expression with distant metastasis in Cancer. They have commented that elevated MIB 1 labeling index may induce epithelialto-mesenchymal transition by increasing the expression of vimentin, which enhances cancer-cell invasion and metastases. 24 Information about metastases in patients in present study is not available.
There were some limitations in present study. Being retrospective in nature, no follow-up about patients studied is available. Being a single observer study, the present study did not evaluate inter-observer variability in determining results. Besides, the present study is a single-institution study.

Conclusion
Cases of high (more than 20%) Ki-67 MIB-1 labeling index can be segregated even from among those cases with low histopathological grade, low mitotic index and pTa histopathological stage. Such cases are candidates for multimodality treatment including radiotherapy, chemotherapy etc.
For segerating tumors with better and worse prognosis, it is needed to adopt uniform cut-off percentage of Ki-67 MIB-1 labeling index in all international studies to enable logical comparison of their results.
As a recent systematic review and meta-analysis 23 has stated that significant correlation between high Ki-67 expression and survival outcome are present only in European-American patients, and it does not reach significant level in Asian population, there is need for multiinstitutional follow-up studies in our country to find the exact value of determination of Ki-67 expression.

Source of Funding
No financial support was received for the work within this manuscript.

Source of Funding
Self only for purchasing the antibody to Ki-67 antigen for immunohistochemical analysis.