Incidental prostate cancer diagnosed following a transurethral resection of the prostate: A national database analysis in England

Abstract

Objectives:

The aim was to compare prostate cancer and all-cause mortality in patients diagnosed with prostate cancer following a transurethral resection of the prostate (TURP) (incidental prostate cancer, IPC), to men diagnosed with localised non-incidental prostate cancer (NIPC).

Patients and methods:

Men diagnosed with localised prostate cancer between 2000 and 2008 were identified from the English national cancer registry. Their records were linked to the Hospital Episode Statistics (HES) database of hospital admissions in England to identify men who had a TURP. Men were considered to have IPC if prostate cancer was diagnosed less than 60 days after TURP. Mortality rates were calculated using the Kaplan–Meier method. Mortality rate ratios (RR) comparing IPC and NIPC were calculated with multivariable Poisson regression adjusting for age group, co-morbidities, year of diagnosis and radical treatment.

Results:

A total of 192,960 men were included. Of these, 6666 (3.5%) had IPC and 186,294 (96.5%) NIPC. Median follow-up was 4.7 years (0.5–11.0). Ten-year prostate cancer mortality was 17.1% in IPC, 19.0% in NIPC. With adjustment, the prostate cancer-specific mortality in IPC was 30% lower than NIPC (RR 0.70, 95% CI 0.65–0.75, p<0.001), with no difference in all-cause mortality.

Conclusion:

Ten-year prostate cancer mortality in IPC was 17%. Men with IPC had lower prostate cancer-specific mortality than other prostate cancer patients.

Keywords

Incidental prostate cancer prostate cancer prostate cancer management transurethral prostatectomy

Introduction

The prognosis of men diagnosed with prostate cancer following a transurethral resection of the prostate (TURP), often referred to as incidental prostate cancer (IPC), is uncertain. There is some evidence from studies carried out in the 1980s that men with IPC are likely to have low rates of disease progression,^1,2 whereas later studies reported significant progression risks.^3–13

A TURP removes tissue mainly from the transition and central zones of the prostate. In men with prostate cancer undergoing a TURP, there is a risk that there is cancer tissue in the residual part of the prostate that is not removed. The observed probability of finding no residual cancer at radical prostatectomy in men with IPC ranges from as little as 2% to as high as 48%.¹⁴ A number of studies have tried to identify factors that predict residual cancer in men with IPC in order to help identify those who may benefit from further radical treatment, but these factors have yet to be validated externally.^14,15 As a result, there is no consensus on the management of men with IPC. European guidelines consider radical treatment to be an option for men with IPC who have a life expectancy of >10 years.¹⁶

Many studies of the prognosis of men with IPC are small and retrospective and were carried out in a single institution when routine prostate-specific antigen (PSA) testing was not used. In this study, we used linked national databases to compare the 10-year prostate cancer-specific and all-cause mortality in English men diagnosed after 2000 with IPC and with non-incidental prostate cancer (NIPC), adjusting for differences in their characteristics and whether or not they underwent radical treatment. This study is part of a programme of work assessing the value of procedure-specific and disease-specific metrics derived from English hospital admission records to assess the performance of English National Health Service (NHS) providers.

Patients and methods

We used English cancer registry data to identify men more than 45 years of age who were diagnosed with prostate cancer between 2000 and 2008.¹⁷ The records of these men were linked to the Hospital Episode Statistics (HES) database, which contains records of all admissions to hospitals of the English National Health Service. In the HES database, diagnoses and cause of death are coded according to International Classification of Disease-10 (ICD-10) codes,¹⁸ and procedures according to the Office of Population Censuses and Surveys Classification of Interventions and Procedures (OPCS-4) codes.¹⁹ Only the primary cause of death was considered. An overview of the procedure codes used is given in Appendix 1.

The linkage of the cancer registry and HES records was based on a hierarchical deterministic approach, which involved matching patient records using the patient identifiers, including the patients’ unique NHS number, date of birth and postcode.²⁰

Figure 1 shows the construction of the study population. A total of 261,622 men aged 45 years or older could be included. Of these men, 252,056 (96.3%) could be linked to HES data. Men who had a record of stage 3/4 disease or metastases and those who died within 6 months of diagnosis were excluded. Using HES data, men were considered to have IPC if they had a recorded TURP less than 60 days before the date of the prostate cancer diagnosis; if they did not have a prostate cancer diagnosis code in the TURP record to ensure that the TURP was undertaken for benign disease; and if they did not have a record of a transrectal (TRUS) or transperineal (TP) prostate biopsy at any time before or less than 60 days after the TURP in HES data.

Co-morbidities were derived from the Royal College of Surgeons of England Charlson Co-morbidity Score.²¹ This validated score included the number of co-morbidities recorded in HES in records of admissions in the year before and at the time of the prostate cancer diagnosis. The cancer registry data were used to identify men who had radical radiotherapy treatment and HES data were used to identify those who had a radical prostatectomy.

Mortality rates were calculated using the Kaplan–Meier method. Prostate cancer-specific mortality was estimated handling death due to other causes as a censoring event. Multivariable Poisson regression analysis was used to estimate mortality rate ratios (RRs), adjusting for age, year of diagnosis, number of co-morbidities and whether or not patients had a radical treatment.

All statistical analyses were conducted using Stata version 11.2 (Stata Corp. 2009. Stata Statistical Software: Release 11, College Station, TX: StataCorp LP).

Results

A total of 192,960 men were included in the study. 6,666 (3.5%) had IPC, 186,294 (96.5%) NIPC. Median follow-up was 4.7 years (0.5–11.0). Men with IPC tended to be older, have fewer co-morbidities, and undergo less often radical prostate cancer treatment (Table 1).

Table 1.

Characteristics of the men included in the study.

Characteristic	NIPC (n, %)	IPC (n, %)	Total (n, %)
Age
45–54	5849 (3.1%)	79 (1.2%)	5928 (3.1%)
55–64	39 656 (21.3%)	818 (12.3%)	40,474 (21.0%)
65–74	74,818 (40.2%)	2374 (35.6%)	77,192 (40.0%)
75–84	54,996 (29.5%)	2745 (41.2%)	57,741 (29.9%)
⩾85	10,975 (5.9%)	650 (9.8%)	11,625 (6.0%)
Number of co-morbidities
0	107,483 (57.7%)	4501 (67.5%)	111,984 (58.0%)
1	60,997 (32.7%)	1576 (23.6%)	62,573 (32.4%)
2	13,454 (7.2%)	450 (6.8%)	13,904 (7.2%)
3	4360 (2.3%)	139 (2.1%)	4499 (2.3%)
Radical treatment
Yes	42,021 (22.6%)	388 (5.8%)	42,409 (22.0%)
No	144,273 (77.4%)	6278 (94.2%)	150,551 (78.0%)
Year of diagnosis
2000	16,111 (8.6%)	820 (12.3%)	16,931 (8.8%)
2001	18,816 (10.1%)	839 (12.6%)	19,655 (10.2%)
2002	19,304 (10.4%)	718 (10.8%)	20,022 (10.4%)
2003	19,729 (10.6%)	705 (10.6%)	20,434 (10.6%)
2004	22,362 (12.0%)	659 (9.9%)	23,021 (11.9%)
2005	21,587 (11.6%)	714 (10.7%)	22,301 (11.6%)
2006	22,951 (12.3%)	692 (10.4%)	23,643 (12.3%)
2007	22,848 (12.3%)	792 (11.9%)	23,640 (12.3%)
2008	22,586 (12.1%)	727 (10.9%)	23,313 (12.1%)
Total	186,294 (96.5%)	6666 (3.5%)	192,960

IPC: incidental prostate cancer; NIPC: non-incidental prostate cancer.

Figure 2 demonstrates that the 10-year prostate cancer-specific mortality in IPC was 17.1% in men with IPC and 19.0% in men with NIPC (unadjusted RR 0.87, 95% confidence interval 0.81–0.94, p<0.001). All-cause mortality was 49.2% in men with IPC and 40.1% in men with NIPC (unadjusted RR 1.27, 1.22–1.33, p<0.001).

With adjustment for patient characteristics (age, year-of-diagnosis and co-morbidities) and whether or not radical treatment was received, the difference in prostate cancer-specific mortality increased (adjusted RR 0.70, 95% confidence interval 0.65–0.75, p<0.001) but the difference in all-cause mortality disappeared (RR 1.01, 95% confidence interval 0.97–1.05, p=0.7).

Discussion

Our population-based study of patients diagnosed after 2000 showed that men with IPC had a 30% lower prostate cancer-specific mortality than men with NIPC when differences in patient characteristics and prostate cancer treatment were taken into account; however, we did not find differences in overall mortality.

Only two other population-based studies have investigated mortality in men with IPC. In Sweden, 10-year prostate cancer-specific mortality, handling death due to other causes as a competing risk, was found to be 26.6% in men diagnosed with IPC and 40.7% in men diagnosed with NIPC between 1970 and 2003.¹³ Their corresponding figures for the PSA era only (men diagnosed between 1992 and 2003) were 23.2% with IPC and 35.1% with NIPC.¹³ The NIPC mortality rate in the Swedish study is higher than we report for England, most likely because the Swedish study included men of all stages whereas we aimed to limit the population included in our study to men with localised prostate cancer. In a US-based study of men diagnosed with IPC between 1992 and 2002 according to linked SEER-Medicare data, 10-year prostate cancer-specific mortality, again with death due to a cause other than prostate cancer handled as a competing event, was less than 10% of IPC patients with Gleason 5–7 and about 30% in IPC patients with Gleason 8–10.²²

The most important methodological limitation of our study is that specific staging data differentiating between T1a/b/c cancers was missing in the majority of the patients (about 80%). For that reason, we identified men with IPC based on whether or not they had a transurethral resection less than 60 days before the recorded diagnosis rather than using a record of stage of T1a (cancer found incidentally during a TURP with cancer in less than 5% of the resected tissue) or T1b (similar definition but with cancer in more than 5% of the resected tissue) in the cancer registry data.

For the same reason, some men with advanced disease may have been inadvertently included in our study population. However, it is estimated that at least 80% of men with prostate cancer have localised disease at the time of diagnoses.²³ Therefore, we believe that only a small proportion of men had advanced disease in this study population. In addition, we tried to further mitigate the impact of differences in cancer stage by excluding men who died within 6 months of diagnosis.

A key strength of our approach is that more than 97% of all cancer diagnoses are captured in English cancer registry data.²⁴ In addition, more than 95% of these could be linked to a HES record, which underlines the representativeness of our study population.

TURPs do not remove the whole prostate, and prostate cancer may be present in residual prostate tissue. In European guidelines for the management of prostate cancer, it is suggested that in some situations, biopsying the residual prostates in men diagnosed with IPC after a TURP in order to obtain a more representative tissue sample, may be appropriate, especially if men have a life expectancy of more than 10 years.¹⁶ Few studies exist on this topic. A small study including only 22 men, concluded that in 86% of men a biopsy following a TURP did not provide further information, although 14% of men had their disease upgraded.²⁵

It has been shown that comparing results of PSA tests carried out before and after surgery for benign disease can provide additional information about the presence of cancer in the residual parts of the prostate, However these models require external validation before routine clinical use.^15,16 Alternatively, multi-parametric magnetic resonance imaging (mp-MRI) may provide further information on extent of prostate cancer following a TURP, but again, few studies evaluate the utility of mp-MRI specifically in this context.²⁶ Also, mp-MRI can be used to detect and, if present, localise cancer tissue in the prostate before a TURP is carried out.²⁷ Although we found that prostate cancer-specific mortality in men with IPC is lower than in men with NIPC, we feel that this type of additional information should be taken into account when treatment decision are made.

Conclusions

We have a found the 10-year prostate cancer mortality in men with IPC is 17%. However, after multivariable adjustment, IPC still had 30% reduced mortality rate when compared to localised non-IPC. IPC is a variable disease. Men with IPC may benefit from further prostate biopsies or imaging of the prostate to exclude disease in the residual prostate.

Footnotes

Appendix 1: Procedure codes used for the analysis of Hospital Episode Statistics (HES) data

OPCS 4.6 codes used in HES data:

Transurethral resection of prostate (TURP):

M65.1 – Endoscopic resection of prostate using electrotome

M65.3 – Endoscopic resection of prostate NEC

M65.8 – Other specified endoscopic resection of outlet of male

bladder

M65.9 – Unspecified endoscopic resection of outlet of male

bladder

M66.2 – Endoscopic incision of outlet of male bladder NEC

M67.8 – Other specified other therapeutic endoscopic operations

on prostate

M67.9 – Unspecified other therapeutic endoscopic operations on

Prostate

Transrectal prostate biopsy:

M70.3 – Rectal needle biopsy of prostate

Transperineal prostate biopsy:

M70.2 – Perineal needle biopsy of prostate

Radical prostatectomy:

M61.1 – Total excision of prostate and capsule of prostate

M61.2 – Retropubic prostatectomy

M61.3 – Transvesical prostatectomy

M61.4 – Perineal prostatectomy

M61.8 – Other specified open excision of prostate

M61.9 – Unspecified open excision of prostate

M62.8 – Other specified other open operations on prostate

M62.9 – Unspecified other open operations on prostate

Acknowledgements

EA received an Academic Clinical Fellowship from the National Institute of Health Research. ME receives research support from the United Kingdom’s National Institute of Health Research UCLH/UCL Biomedical Research Centre. We would like to thank Amundeep Johal and Paul Cathcart for their assistance in accessing the administrative data and in data cleaning.

Conflicting interests

EA and JvdM declare no competing interests. ME receives funding from Sonacare Inc, GSK and Advanced Medical Diagnostics for clinical trials. ME was a previously paid consultant to Steba Biotech, GE Healthcare/Oncura and Astra-Zeneca. ME is a paid adviser to Nuada Medical Ltd. None of the funding sources had any role in the data acquisition, analyses and production of this manuscript.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Informed consent

Informed consent was not sought for the present study because this study is exempt from UK National Research Ethics Committee approval as it involved the analysis of an existing database of anonymised data for service evaluation.

Ethical approval

This study is exempt from UK National Research Ethics Committee approval as it involved the analysis of an existing database of anonymised data for service evaluation.

Guarantor

EA.

Contributorship

ME and JvdM conceived the idea of the study. EA and JvdM were involved with the study design and with data analysis. The first draft of the manuscript was written by EA. All authors reviewed and edited the manuscript and approved the final version of the manuscript.

References

Cantrell

DeKlerk

Eggleston

. Pathological factors that influence prognosis in stage A prostatic cancer: the influence of extent versus grade. J Urol 1981; 125: 516–520.

Humphrey

Vollmer

. The ratio of prostate chips with cancer: a new measure of tumor extent and its relationship to grade and prognosis. Human Pathol 1988; 19: 411–418.

Tombal

De Visccher

Cosyns

. Assessing the risk of unsuspected prostate cancer in patients with benign prostatic hypertrophy: A 13-year retrospective study of the incidence and natural history of T1a-T1b prostate cancers. BJU Int 1999; 84: 1015–1020.

Cheng

Bergstralh

Scherer

. Predictors of cancer progression in T1a prostate adenocarcinoma. Cancer 1999; 85: 1300–1304.

Amakasu

Akimoto

Akakura

. Disease progression in stage A prostate cancer. Int J Urol 1995; 2: 39–43.

Brawn

Johnson

Speights

. Long-term survival of stage A prostate carcinoma, atypical hyperplasia/adenosis and BPH. Br J Cancer 1994; 69: 1098–1101.

Thompson

Zeidman

. Extended follow-up of stage A1 carcinoma of prostate. Urology 1989; 33: 455–458.

Blute

Zincke

Farrow

. Long-term followup of young patients with stage A adenocarcinoma of the prostate. J Urol 1986; 136: 840–843.

Epstein

Walsh

Brendler

. Radical prostatectomy for impalpable prostate cancer: the Johns Hopkins experience with tumors found on transurethral resection (stages T1A and T1B) and on needle biopsy (stage T1C). J Urol 1994; 152(Pt 2): 1721–1729.

10.

Roy

Horne

Raife

. Incidental carcinoma of prostate, Long-term follow-up. Urology 1990; 36: 210–213.

11.

Beynon

Busuttil

Newsam

. Incidental carcinoma of the prostate: selection for deferred treatment. Br J Urol 1983; 55: 733–736.

12.

Epstein

Paull

Eggleston

. Prognosis of untreated stage A1 prostatic carcinoma: a study of 94 cases with extended followup. J Urol 1986; 136: 837–839.

13.

Andrèn

Garmo

Mucci

. Incidence and mortality of incidental prostate cancer: a Swedish register-based study. Br J Cancer 2009; 100: 170–173.

14.

Capitanio

Briganti

Suardi

. When should we expect no residual tumor (pT0) once we submit incidental T1a-b prostate cancers to radical prostatectomy? Int J Urol 2011; 18: 148–153.

15.

Capitanio

Scattoni

Freschi

. Radical prostatectomy for incidental (stage T1a-T1b) prostate cancer: analysis of predictors for residual disease and biochemical recurrence. Eur Urol 2008; 54: 118–125.

16.

Heidenreich

Bastian

Bolla

EAU. Guidelines on Prostate Cancer, http://uroweb.org/wp-content/uploads/09_Prostate_Cancer_LR.pdf (2013, accessed 13 Aug 2013).

17.

National Cancer Intelligence Network (NCIN). Merged English Cancer Registry Data, http://www.ncin.org.uk/collecting_and_using_data/national_cancer_data_repository/cancer_registry (2010, accessed 3 Aug 2013)

18.

World Health Organization. ICD-10 Version: 2010, http://apps.who.int/classifications/icd10/browse/2010/en (2010, accessed 2 Sep 2013)

19.

NHS Connecting for Health. OPCS Classification of Interventions and Procedures Version 4.6 (April 2011) Volume 1 - Tabular List. London: The Stationery Office; 2011.

20.

Health and Social Care Information Centre. Data Linkage and Extract Service, http://www.hscic.gov.uk/dles (2013, accessed 12 Jan 2014)

21.

Nuttall

Van der Meulen

Emberton

. Charlson scores based on ICD-10 administrative data were valid in assessing comorbidity in patients undergoing urological cancer surgery. J Clin Epidemiol 2006; 59: 265–273.

22.

Lu-Yao

Albertsen

Moore

. Outcomes of localized prostate cancer following conservative management. JAMA 2009; 302: 1202–1209.

23.

National Cancer Institue. Cancer of the Prostate - SEER Stat Fact Sheets, http://seer.cancer.gov/statfacts/html/prost.html (2010, accessed 28 Sep 2013)

24.

Office for National Statistics. Cancer incidence and mortality in the United Kingdom. Newport: 2013.

25.

Lee

Thiruneelakandasivam

Hong

. Are transrectal prostate biopsies routinely indicated in patients with incidentally diagnosed prostate cancer following transurethral resection of the prostate for benign disease? Urologia Internationalis. 2013; 91: 397–403.

26.

Sheu

Chiang

Wang

. Transurethral resection of the prostate-related changes in the prostate gland: correlation of MRI and histopathology. J Comput Assist Tomogr 2000; 24: 596–599.

27.

Barentsz

Richenberg

Clements

. ESUR prostate MR guidelines 2012. Eur Radiol 2012; 22: 746–757.