SEARCH

SEARCH BY CITATION

Keywords:

  • Prostate cancer;
  • localized;
  • incidental;
  • natural history;
  • biology;
  • follow-up

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References

Objective

To determine the natural course of incidental untreated transition zone prostate cancer and thus help to identify criteria to predict the prognosis and to determine treatment for individual patients.

Materials and methods

A total of 1135 unselected surgical specimens of the prostate, examined during 1974 and 1975, were reviewed while unaware of case by two experienced pathologists. The patients from which the samples were obtained were followed for up to 20 years or death by The Cancer Registry of Norway and the outcome compared with the histological review.

Results

The histology review revealed a total of 311 cancers, of which 73 had not been initially recorded; these patients had received no treatment. The kappa coefficient for interobserver reproducibility was 0.86 for carcinoma. The follow-up showed that patient age was the strongest predictor of survival, followed by histological grade and percentage of tumour involvement. Only two of the 73 patients with untreated transition zone cancer died from prostate cancer during the follow-up, compared with 78 of 144 patients with standard management of transition zone tumours. The 5- and 10-year relative survival rates for the 144 patients with standard management of transitional zone tumours and for the 53 patients with peripheral zone tumours were 56% and 26%, and 45% and 33%, respectively. Metastasis (+ or −) was the only individual prognostic factor in the multivariate analysis.

Conclusion

This study shows that patients with incidental low-grade tumours have a low probability of dying from prostate cancer and may thus be followed expectantly. The biological distinction between atypical hyperplasia and stage T1a cancer is unclear. The survival of men with prostate cancer is significantly reduced with loss of differentiation and with increasing tumour volvement.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References

Prostate cancer is unique among potentially lethal human malignancies in having a wide discrepancy between the high prevalence of histological changes recognizable as cancer and the much lower prevalence of clinical disease. Only 20% of all prostate cancers are considered to be clinically important [1]. Localized prostate cancer may be curable by total prostatectomy or definitive radiotherapy, but no trial has documented a possible benefit of radical treatment of early cancer. Overtreatment of naturally indolent disease may result in needless morbidity and mortality, and consumption of scarce resources. There is presently no reliable biomarker to predict the biological outcome and guide the clinician in the selection of the appropriate treatment. Assessment of the likely course of untreated prostate cancer is therefore important.

Tissues obtained during TURP and transvesical enucleation (TV) are mostly derived from the transition zone, whereas core needle biopsies (CNBs) are derived mainly from the peripheral zone [2[3]–4]. Most prostate cancers originate in the peripheral zone; transition zone cancers constitute a small proportion of the total number of tumours and are considered to be less aggressive [3,5]. However, it has been shown that a significant proportion of transition zone cancers pursue an aggressive course [6]. Thus it is important that these are detected; the aim of the present study was to determine the natural course of incidental untreated transition zone prostate cancer.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References

The present study was based on all available material from prostatic resections and core biopsies from 1135 patients studied at the Department of Pathology, Ullevaal Hospital, Oslo in 1974 and 1975. All patients were seen as in- or outpatients at two main urological departments in Oslo, and there was no known selection bias inherent in the case series.

Overall, 197 cases of prostate cancer were diagnosed in 1974–75, of which 144 were detected in TURP and TV specimens, and 53 in CNB specimens. At the time of the start of the study (1974–75), treatment of prostate cancer in Norway generally consisted of surgical or medical castration (the latter by oestrogen) or no treatment, and we presume that these patients had been managed accordingly. Histological re-examination in 1995 revealed 73 additional cases of cancer in TURP and TV specimens that were not previously registered in the Cancer Registry of Norway (CRN); 39 cases had not been diagnosed at the original examination of the slides in 1974–75, in most cases because stricter diagnostic criteria were applied at re-examination in 1995. In 34 cases the primary diagnosis was ‘focus of atypical small alveolar proliferation’. In the present study, the 73 tumours diagnosed at re-examination in 1995 are recorded as untreated and the 197 tumours diagnosed in 1974–75 as treatment by ‘standard management’ (SM).

The total material comprising 79 CNB, 731 TURP and 325 TV specimens had been processed as follows. For CNB, all tissue had been embedded; TURP chips were mixed well and three to four paraffin blocks embedded. When cancer was suspected, all tissue was embedded; the TV specimens consisted of nodules that were transected. Three to four paraffin blocks from the largest nodules were embedded. Details of the material and methods were published elsewhere [7].

The 1135 specimens were reviewed while unaware of case origin by two pathologists (F.S. and A.B.) and the following variables recorded; type of material (TURP, CNB, TV); histological grade of cancer as defined by the WHO criteria; prostatic intraepithelial neoplasia (PIN) and atypical adenomatous hyperplasia (AAH); the extent (area) of tumour tissue on the slide (0–5%, 6–25%, 26–50%, 51–75%, >75%) was noted on TURP and TV specimens. In cases where there was diagnostic disagreement between the pathologists, a consensus diagnosis was reached using a double-view microscope. The patients were followed for up to 20 years or death by the CRN; the median follow-up was 11 years. Cancer reporting in Norway has been compulsory by law since 1952 and the completeness of prostate cancer reporting in the CRN is >99% [8]. The following variables were registered by the CRN: type of cancer, metastasis (+/−) follow-up status (dead/alive) and the date of death, given in collaboration with Central Bureau of Statistics. The TNM classification system was not widely used in Norway in 1975 and TNM data for analysis were not available.

The influence of age at diagnosis, histological grade, extent of tumour involvement, metastatic status (+/−) and PIN was analysed using a Cox proportional-hazards survival model. Relative survival rates by age groups were calculated for treated and untreated groups of transition zone carcinomas and for peripheral zone carcinomas. This implies that observed survival is adjusted for expected survival based on the life tables of the Norwegian population. The standard deviation of the relative survival rates was computed according to methods described by Greenwood [9].

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References

No patients were lost to follow-up. The overall agreement between the pathologists for cancer diagnosis using the WHO grading system was 83% and the weighted kappa coefficient was 0.86. On re-examination of the 1135 specimens, 270 cancer cases not registered in the CRN before 1974 were found; 217 cancers were transition zone tumours (175 TURP and 42 TV specimens) and 53 were peripheral tumours diagnosed in CNB specimens. The mean number of tissue blocks for TURP and TV specimens was 4.3 (range 2–10). Examination of the files of the CRN disclosed that only two (3%) of the 73 presumed untreated patients with transition zone cancer were recorded as dying from prostate cancer, after 2 and 8 years. There were 24 (17%) and 19 (26%) patients with a second malignant tumour in the SM and untreated population, respectively.

Table 1 shows the main histopathological findings in the transition zone tumours; most (69%) of the untreated tumours were well differentiated and 77% involved <5% of the tissue specimen in the PT1a category [9]. Most of the SM tumours (54%) were moderately differentiated and only 31% were small, involving <25% of the specimen area. The peripheral zone cancers had almost the same distribution of histological grade as the SM tumours (data not shown).

Table 1.  Observed survival (in months) of 217 patients with transition zone tumours related to grade (WHO) and cancer involvement. Group A comprised 73 untreated patients and group B 144 patients undergoing standard management Thumbnail image of

As shown in Table 1, the median survival of the untreated patients was almost twice that of the SM patients, at 83.0 and 47.5 months, respectively. Furthermore, survival significantly increased in patients with well-differentiated tumours. There was a trend towards reduced survival in patients with increasing amount of tumour tissue in the untreated group, but the difference was not statistically significant.

The 5- and 10-year relative survival rates are given in Table 2. The 5-year relative survival rate for the untreated and the SM patients with transition zone tumours was 92% and 56%, respectively. Whereas the values for the untreated patients seemed to stabilize at 5 years, the corresponding 10-year relative survival rate for the SM patients decreased to 26%. There were only minor differences in survival rates between the SM transition zone tumours and patients with peripheral zone tumours.

Table 2.  The 5- and 10-year relative survival related to patient age and median relative survival. Group A comprised 73 untreated transition zone carcinomas, group B 144 transition zone carcinomas with standard management and group C 53 peripheral carcinomas Thumbnail image of

Eighty of the 217 patients were recorded as dying from prostate cancer during the follow-up and all except two were in the SM group. Of the 78 cancer deaths in this group, 39 and 35 were poorly and moderately differentiated. Furthermore, 48 SM patients had large tumours involving >50% of the specimen area (>75%, 33; 51–75%, 15) and only five were T1a (<5%).

Table 2 also shows that cancer-specific survival was age-related; in all subgroups, patients <65 years old survived significantly longer than patients in the higher age groups. On the other hand, there were only minor differences in patient age at diagnosis for the various categories; the mean ages of patients with transition zone carcinomas (73 untreated and 144 SM) and 275 patients with BPH (with no AAH or PIN) were 71.0 (69.1–72.9), 73.2 (71.9–74.2) and 69.3 (68.3–70.3) years, respectively. Thus for patients with untreated cancer the mean age was 2.2 years lower than for SM patients. Patients that were still alive after the maximum follow-up were 8.5 years younger than the control group without cancer. In a multivariate analysis, only metastatic status was statistically significant (P<0.001).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References

Concern has been raised that development in early diagnosis and increasing public awareness of prostate cancer will identify many incidental tumours that pose no threat to a patient’s life [10]. Thus, the single most important question is how to distinguish tumours that will progress from those which will not during the expected lifetime of the patient. By re-examining biopsies from a defined urological population, we identified 73 patients with incidental prostate cancer followed for up to 20 years with no treatment. There was no known bias inherent in the selection of patients.

Current staging of early prostate cancer allocates patients according to the method of tumour detection, separating impalpable transition zone cancers detected during TURP and TV (T1) from palpable peripheral zone cancers verified in CNB specimens (T2). Incidental adenocarcinomas found on TURP and TV are divided into stages T1a and T1b [11]. Data on clinical staging were not available in the present study and 5% tumour involvement was used as the threshold to identify T1a tumours, as recommended by the TNM staging system [11]. Generally, transition zone tumours appear to be smaller and better differentiated than peripheral zone tumours [3,4,12,13].

Central to the present study is that 73 histological cancers were overlooked in the original histopathological assessment in 1974–75; 35 of these 73 tumours were described as benign, with irregular and/or atypical epithelium, and 38 were reported as benign without further comments (data not shown). It may be difficult even among experienced histopathologists to distinguish irregular and/or atypical glandular proliferation from adenocarcinoma. However, a weighted kappa coefficient showed good agreement (0.86) between the pathologists, who also participated in the diagnostic evaluation of these patients in 1974–75. The 35 cases with irregular and/or atypical epithelium indicate that stricter malignancy criteria were used in 1994–95. These tumours clearly represent a selected group of small, well differentiated tumours (Table 1) that may easily be overlooked or misinterpreted.

Relatively few studies have reported on the natural history of incidental prostate cancer and they indicate that stage T1a and T1b tumours have a significantly different prognosis [14[15][16][17]–18]. Only 4% of patients with clinical stage T1a disease have been reported to develop disease progression 4 years after diagnosis, with 16–25% showing progression 8–10 years after TURP [14[15]–16]. In contrast, ≈35% of men with stage T1b disease showed tumour progression at 4 years [15] and 53% at 8 years [17]. The present report confirms that T1a tumours are less aggressive than T1b tumours. Patients with SM and untreated T1a tumours survived significantly longer than patients with T1b tumours (mean 36 months, data not shown). We also noted a 20-year progression rate of 3% among the untreated patients, compared with a 10% progression rate reported by Zhang et al. [17]. The good outcome of T1a tumours may be biased by the possibility that all tumour tissue was resected by TURP and TV [3,4,18].

Recent reports of behaviour of patients with clinically stage T1–2 prostate cancer show that progression occurs slowly but relentlessly in the absence of treatment [19[20][21][22][23][24][25][26][27][28]–29], with local recurrence often preceding distant metastases [22,25]. A pooled analysis of six large individual studies by Chodak et al. [30] concluded that tumour grade was the only individual predictor of cancer-specific mortality. At diagnosis, 55–60% of cancers are localized, comprising patients with stage T1–2 tumours [1]. The TNM data were not available in the present study and thus the data are not comparable with those cited. Histological grade and extent of tumour tissue predicted patient outcome in the univariate analysis, in agreement with other studies [22,25,28[29]–30], but in the multivariate analysis, metastatic status was the only factor that significantly predicted outcome. The favourable outcome of the untreated population reflects that most of these tumours were well differentiated, involving <25% of the specimens. Surprisingly, the relative survival rates of the SM patients was similar to that of those with transitional and peripheral zone tumours, which contradicts the current hypothesis that peripheral zone tumours are more aggressive [3].

It is possible that the two patients with untreated cancer who were reported as dead from prostate cancer after 2 and 8 years may have had predominantly peripheral zone tumours, with limited invasion of the transition zone, thus producing only a small amount of tumour tissue in the TURP specimens.

It is questionable whether any form of treatment favourably alters the natural history of prostate cancer. In a literature review of the results of radical compared with deferred treatment of localized prostate cancer, the corrected survival rates after 10 years were 93%, 84% and 74%, respectively, for radical surgery, deferred treatment and radiation therapy [23]. On the other hand, the recent results of prostate carcinoma screening in the USA [10] has shown a 100% 10-year relative survival rate for patients treated with prostatectomy. Radical surgery was not performed in Oslo during the present study period and most patients received deferred treatment. It is thus less likely that the current study is biased by choice of treatment.

The published data on associations between patient age at diagnosis and survival of prostate cancer are conflicting [10,20,29[30][31][32][33]–34]. In the current study, the mean age at diagnosis of untreated patients was 71 years, compared with 73.4 and 73.0 for SM patients treated for cancer and those dead from cancer, respectively. Patient age in the control group was significantly lower (mean 4.1 years) that of the SM population. The surprising low age of patients still alive (mean 60.8 years) may be influenced by the high frequency of low-grade and low-stage tumours in these patients, who were all in the untreated group. The occurrence of comorbidity is usually low in younger age groups [35] and Chodak et al. [30] also reported that being younger than 61 years had a significant positive effect on survival, in agreement with the present report (Table 2).

The high percentage of incidental cancer in the transition zone probably arises through selective sampling by TURP and TV. Up to 30% of stage T1 tumours could theoretically be removed [3,4,12,13,18]. Numerous studies have investigated the minimum number of blocks that need to be submitted to accurately detect incidental adenocarcinoma in TURP specimens; the number of blocks in the present series was in accord with these recommendations [36]. When stage T1a carcinoma was suspected, all tissues were embedded, in agreement with the recommendations of McDowell et al. [37].

In conclusion, the results of this follow-up study may have some important biological and clinical implications. Nearly all (97%) patients with untreated prostate cancer had incidental cancer [11]. Most of these were stage T1a, initially often reported as irregular small alveolar or atypical prostatic hyperplasia. Although pathologists attempt to distinguish between irregular and/or atypical hyperplasia and cancer, the biological distinction is unclear. It is important that the clinician is aware of the relatively benign outcome of these borderline tumours, to avoid overtreatment.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References