Prognostic factors associated with bladder tumours

Introduction: Transurethrally resected bladder tumours are at risk of recurrence and progression to invasive cancer. Objectives: To assess factors which predict recurrence and progression in bladder tumours. Methods: This retrospective analysis included 192 patients with bladder tumours treated at a urology unit in a tertiary care hospital in Sri Lanka with a mean follow up of 63.7 months (SD ± 18.8, range 37-99 months). Follow up details were obtained from clinic records. Factors which were analyzed as possible predictors of tumour recurrence and progression included the tumour growth pattern, mitotic count, necrosis, lamina propria invasion, muscularis propria invasion, lympho-vascular invasion, focal pleomorphic areas, squamous differentiation, glandular differentiation, adjacent carcinoma in-situ and the tumour grade. Cox univariate and multivariate analysis was done together with a Kaplan Meier survival analysis. Results and Conclusion: In the univariate analysis the tumour stage (p=0.042) and lamina propria invasion (p=0.031) were the only significant predictors of tumour recurrence. In multivariate analysis the most significant independent factor associated with tumour recurrence was lamina propria invasion (p=0.02), In Kaplan Meier survival analysis there was a significant difference in recurrence free survival (RFS) between the low grade urothelial carcinoma and the invasive group within the WHO/ISUP classification (log rank 4.78, p=0.0287). Only six cases progressed in stage or grade during the follow up period. Journal of Diagnostic Pathology 2013 (8); 1:34-49


Introduction
Bladder cancers account for 1% of all cancers in Sri Lanka (1).Transurethrally resected bladder tumours are at risk of recurrence and progression to invasive cancer (2).Therefore, information is necessary to identify patients at risk of recurrence and progression.Identifying the prognostic factors that determine the risk in each patient for recurrence and progression remains a subject of extensive research.
Prognostic factors which predict tumour recurrence and progression have not been studied previously in Sri Lanka.Thus the objective was to determine factors which predict recurrence and progression of bladder tumours in a local setting.

Methods
A retrospective analysis of histological assessment performed on 262 patients diagnosed with bladder tumours who underwent surgery in a urology unit at a tertiary care hospital in Sri Lanka from September 1996 to December 2001 was done.Of these, 50 cases with a previous diagnosis of bladder tumour at presentation and 20 cases without demographic data were excluded from the study and the final study population consisted of 192 cases with primary bladder tumour.
Out of these 192 cases, 87.6 % (n=169) were males and 12.4% (n=23) were females (male: female ratio of 7:1).The mean age of the study group was 62 years (SD ± 12.6, range 26 to 91).These cases were followed up for tumour recurrence and progression until December 2004.The mean follow up period was 63.7 months (SD ± 18.8, range 37 to 99 months).

Mitotic activity was assessed as follows:
The mitoses were counted in the most cellular and pleomorphic areas of the tumours using an Olympus (model BX50) conference microscope.
The count was obtained in four sets of 10 consecutive high power fields (x400) and the mean mitotic count was calculated for each tumour.These were further sub classified to mitotic count of ≥ 5/10 HPF and <5/10 HPF.
The presence of highly atypical areas admixed with low grade malignancy was defined as atypical areas containing cells with nuclear grade 3 or 4 (with hyperchromasia and pleomorphism) with or without bizarre mitoses, clearly seen separately from the surrounding tissue.This atypia had to be recognizable under medium power (x 100) The presence of lamina propria involvement was further subdivided into focal and diffuse involvement.One or two foci of lamina propria involvement was defined as focal involvement.Multifocal invasion was defined as diffuse involvement.
The presence of muscle involvement was defined as the invasion the muscularis propria by the tumour.Only thick bundles of smooth muscle present in the biopsy were considered as muscularis propria.Involvement  The findings in the univariate analysis are shown in table 3.In the univariate analysis the only significant predictors of tumour recurrence were lamina propria invasion (HR=2.22,p=0.031,CI 1.075, 4.584) and tumour stage (p=0.042).A significant differencein tumour stage was seen only between pTatumours and pT1 tumours (HR=2.61,0.015, CI 1.207, 5.642).The pattern of invasion of lamina propria, whether focal or diffuse did not significantly affect the recurrence (HR= 2.113, p = 0.059, CI 0.971, 4.599).Muscle invasion was not a significant predictor of recurrence (HR= 0.87, p=0.69).The grade of the invasive tumour (low or high grade) did not significantly affect the tumour recurrence.

Prognostic factor Growth pattern
In the multivariate analysis, the only variable that was significantly associated with tumour recurrence was lamina propria invasion (p=0.02).The tumours with lamina propria invasion were 2.4 times likely to recur than tumours without lamina propria invasion (CI 1.148, 4.974).
The tumour growth pattern, mitotic count, necrosis, lympho-vascular invasion, muscularis propria invasion, focal pleomorphic areas, squamous differentiation, glandular differentiation, adjacent CIS, tumour stage and the WHO grade, were not significantly associated with tumour recurrence.
Analysis with Kaplan Meier survival analysis also showed a significant difference between the recurrence free survival of tumours with lamina propria invasion and without lamina propria invasion (log rank test=4.97,p=0.0258) (Fig 7).

Progression
Only six patients had progression in stage or grade during the follow up period.

Recurrence versus tumour grade
Kaplan Meier survival analysis comparing the recurrence free survival for different WHO/ISUP grades showed a significant log rank test value for all groups (log rank=4.76, p=0.0291).(Fig. 8).
When different curves were compared with log rank test the only significant difference seen was between the low grade urothelial carcinoma and the invasive group (log rank 4.78, p= 0.0287).Note: Censored cases were cases which were not associated with recurrence.

Discussion
Extensive studies have assessed the prognostic value of different clinico-pathlogical variables as possible predictors of tumour recurrence and progression; eg: tumour stage (3), grade (3), size (3,4,5), multiplicity (3,4,5), location (5), associated CIS (5).Furthermore, several cellular and nuclear markers have been identified as having potential prognostic value (15).Some studies have developed prognostic indexes to predict recurrence (3,5).In this study we used multivariate analysis as opposed to some earlier studies which used univariate analysis.Lamina propria invasion and tumour stage are major factors for recurrence in most studies (3,5).However some studies did not show any significant relationship (4).Most studies only studied pTa and pT1 tumours labeling them as superficial tumours (4,5).Tumour stage (ie: presence or absence of lamina propria invasion) has been a significant predictor of recurrence in some studies (5).The terminology used in some studies as 'superficial bladder tumours' is somewhat confusing.Most clinicians also refer to both non invasive (pTa) and lamina propria invasive (pT1) tumours as 'superficial tumours' further confusing the terminology.We believe that the term 'superficial bladder tumours' should be abandoned and be replaced with non invasive (pTa), lamina propria invasive (pT1) and musculairs propria invasive (pT2) tumours as suggested by the International Society of Urological Pathologists (ISUP) conference committee (11).Tumour grade has not been a significant prognostic factor for recurrence in most studies.
A study done by Oosterius in 2002, showed that there was no significant difference in five year recurrence free survival between different groups categorized under WHO/ISUP grading system.They only found a difference in progression between PUNLMP and high grade urothelial carcinoma.Samaratunga et al revealed that the WHO/ISUP grading system is an independent predictor of tumour progression (12,13,14,15).Multivariate analysis showed that the most powerful predictors of recurrence, progression and tumour related death were the previous recurrence rate, tumour size, tumour grade and positive 3 month cystoscopy (16).
In 2000 Rodriguez analyzed 1529 cases with primary superficial bladder tumours for a median follow up period of 4.2 years.Multivariate analysis revealed that multiple tumours, tumour size> 3 cm, CIS increased whereas BCG instillations decreased the risk of recurrence.
Later he identified three risk groups based on these factors and found a significant difference between the groups for predicting recurrence, progression and mortality (17).
The presence of muscle invasion has not been a predictor of recurrence in our study.However when the log rank test was compared between the groups, the difference was only significant between low grade transitional (noninvasive) cell carcinomas and inasive neoplasms.

Conclusion
The most important predictor of tumour recurrence in our study was lamina propria inva-

Fig 5 .Fig 6 .
Fig 5. Invasive high grade urothelial carcinoma with lamina propria invasion (H & E x 100) There were eight high grade transitional cell carcinomas in the study none of which recurred during the follow up period.Three PUNLMP's were present in the study and one case recurred during the follow up period.A comparison of the recurrence free survival between different categories of WHO/ISUP grade with log rank test and p values is shown in table 4.

Fig 8 .Fig. 7 .
Fig 8.The Kaplan Meier survival plot for recurrence free survival comparing different subgroups of the WHO/ISUP system.The combined log rank test was 4.76 (p= 0.0291).Comparison between the survival curves by log rank test is shown in table 4.Note: Censored cases were cases which were not associated with recurrence

In 1995
Kurth assessed factors affecting recurrence, progression and death from bladder cancer in 576 cases with bladder tumours.Allard et al (1998) described a prognostic index based on the number of adverse primary tumour characteristics ie: Primary tumour multiplicity, diameter >3 cm, stage T1 and grade 2 or 3 (3).Their study included 333 patients with primary Ta and T1 bladder cancer and the mean follow up period was 35.3 months.They found that their simple proposed prognostic index was a strong indicator f the clinical course of superficial bladder cancer within 3 years of the transurethral bladder resection (TURBT), but the relative value of each of these prognostic factors was not considered.Dein B. et al (2003) studied 377 cases of bladder tumours randomizing them into six different groups based on different treatment methods and found that on multivariate analysis tumour stage, DNA ploidy, multiplicity, history of recurrence, tumour configuration and type of adjuvant therapy independently affected the rate of recurrence during a mean follow up period of 58 months.They also developed a prognostic index dividing tumours into three risk categories (5).In our study 25% of the primary bladder tumours showed at least one recurrence during the mean follow up period of 63.7 months.In Paper Journal of Diagnostic Pathology 2013 (8); 1:34-49 the univariate analysis lamina propria invasion (HR= 2.22, p 0.031, CI 1.075, 4.584) and tumour stage (p=0.042)were the only significant predictors of tumour recurrence.For tumour stage this difference was only significant for pTa and pT1 tumours (HR 2.61, p=0.015,CI 1.207, 5.642) further signifying the importance of lamina propria invasion for tumour recurrence.The pattern of invasion of lamina propria whether focal or diffuse, did not significantly affect the recurrence.Muscle invasion was not a significant predictor of recurrence.It is possible that exclusion of 56 cases which lacked the muscularis propria on biopsy resulting in a smaller sample size may have contributed to this negative result.The grade of the invasive tumour (low or high grade) did not significantly affect the recurrence.The tumour growth pattern, mitotic count, necrosis, lymphovascular invasion, focal pleomorphic areas, squamous differentiation, glandular differentiation, adjacent CIS, the WHO/ISUP grade of the tumour were not significant predictors of recurrence.On multivariate analysis the only significant independent risk factor for tumour recurrence was lamina propria invasion.The tumours with lamina propria invasion had a 2.4 times risk of recurrence than tumours without lamina propria invasion.Kaplan Meier survival analysis also shows a significant difference in recurrence free survival for tumours with and without lamina propria invasion (log rank 4.97, p= 0.025).
We feel that subcategorizing tumours into an invasive subgroup in the WHO/ISUP classification is an important step forward when grading the bladder tumours.We could not compare recurrence free survival for different grades of non invasive bladder tumours in the WHO/ISUP system as we had relatively small number of cases of non invasive high grade urothelial carcinoma and PUNLMP.There were only 8 cases of high grade non invasive urothelial carcinoma and none of these cases recurred during the mean follow up period of 67.7 months.Only three cases were categorized as PUNLMP in which only one case recurred.There were no tumours which belonged to the papilloma group.Relatively small number of cases of PUNLMP and high grade urothelial carcinoma made it difficult to interpret the data belonging to these groups.Several limitations should be considered in the present study.As mentioned, the small sample size and the follow up period are some of the factors.The tumour size and multiplicity could not be analyzed as possible predictors of tumour recurrence due to lack of clinical information.As some of our patients were given adjuvant intravesical therapy the true natural history and tumour grade may have been underestimated.Furthermore grading of the recurrent tumours would have been affected by the therapy to some extent.