Abstract
Aims:
Cribriform morphology, which includes intraductal carcinoma (IDCP) and invasive cribriform carcinoma, is an indicator of poor prognosis in prostate cancer. Phosphatase and tensin homologue (PTEN) loss is a predictor of adverse clinical outcomes. The association between PTEN expression and morphological patterns of prostate cancer is unclear.
Methods and results:
We explored the association between PTEN expression by immunohistochemistry, Gleason pattern 4 morphologies, IDCP and biochemical recurrence (BCR) in 163 radical prostatectomy specimens. IDCP was delineated from invasive cribriform carcinoma by p63 positive immunohistochemical staining in basal cells. Combined invasive cribriform carcinoma and IDCP were associated with a higher cumulative incidence of BCR [hazard ratio (HR) = 5.06; 2.21, 11.6, P < 0.001]. When including PTEN loss in the analysis, invasive cribriform carcinoma remained predictive of BCR (HR = 3.72; 1.75, 7.94, P = 0.001), while PTEN loss within invasive cribriform carcinoma did not. Glomeruloid morphology was associated with lower odds of cancer stage pT3 and lower cumulative incidence of BCR (HR = 0.27; 0.088, 0.796, P = 0.018), while PTEN loss within glomeruloid morphology was associated with a higher cumulative incidence of BCR (HR = 4.07; 1.04, 15.9, P = 0.043).
Conclusions:
PTEN loss within glomeruloid pattern was associated with BCR. The presence of any cribriform pattern was associated with BCR, despite PTEN loss not significantly associated with invasive cribriform carcinoma. We speculate that other drivers independent from PTEN loss may contribute to poor prognostic features in cribriform carcinoma.
Keywords: biochemical recurrence, cribriform, intraductal carcinoma, pathology, prostate cancer
Introduction
Prostate cancer is the second most common malignancy in men in the United States, with an estimated 248,530 new cases and 34,130 deaths in 2020.1 Cancer grade is a predictor of clinical outcomes.2–4 Gleason grade 4 prostate cancer includes several morphologies, such as poorly formed, glomeruloid and cribriform patterns. Intraductal carcinoma of the prostate (IDCP) is characterised by tumour spreading within prostatic ducts and acini.2 Recent studies have shown that the cribriform pattern, including IDCP and invasive cribriform carcinoma, is a poor prognostic indicator.5–12 Cribriform pattern detected on prostate biopsy predicts adverse pathology at radical prostatectomy,5–7 and is associated with biochemical recurrence (BCR), progression to metastatic disease and decreased cancer-specific survival8–12 when detected on radical prostatectomy. Furthermore, invasive cribriform carcinoma and/or IDCP are associated with molecular features of aggressive, lethal prostate cancer.13 However, these data are primarily drawn from studies analysing IDCP and invasive cribriform carcinoma in a composite fashion.
Studies investigating the association of cribriform pattern and prognosis in patients with prostate cancer have produced conflicting results. One possible explanation could be differences in phosphatase and tensin homologue (PTEN) loss among different patterns. Loss of the PTEN tumour suppressor gene is one of the most commonly occurring genetic aberrations in prostate cancer.14,15 In patients with prostate cancer, PTEN loss is a predictor of adverse pathological features and worse clinical outcomes.16–18 Previous studies have evaluated the association between PTEN expression and different prostate cancer patterns, including cribriform and fused pattern 4,19 combined invasive cribriform pattern and IDCP20 and IDCP.21 This study evaluated the association between PTEN expression by immunohistochemistry, Gleason grade 4 patterns and IDCP in the same cohort, focusing upon predictors of cancer stage and BCR.
Methods
The study was approved by the Vanderbilt Institutional Review Board (160838, 2/18/22). A retrospective search of the pathology files at Vanderbilt University Medical Center from 2005 to 2016 was conducted for consecutive radical prostatectomy specimens with Gleason pattern 4 prostate cancer, including grade groups 2–5. For grade group 4, only Gleason score 4 + 4 tumours were included. Gleason scores 3 + 5 and 5 + 3 tumours were not included due to their rare occurrence. Analysis of minor tertiary components was not performed in this study. Cribriform pattern was defined as a solid proliferation with multiple ‘punched out’ lumina and without intervening stroma.11 Distinction between small and large cribriform patterns was not performed in this study due to the lack of uniform diagnostic criteria and definitive prognostic significance.2 Glomeruloid pattern was defined as dilated glands containing intraluminal solid or cribriform structures with one attachment to the outer epithelial lining and separated from it by a space > 50% of the circumference, resembling a renal glomerulus.22 Bridging with the outer lining was not permitted. When attachment was not identified, the presence of a space between the intraluminal proliferation and the lining epithelium of the outer gland, as defined above, was considered sufficient for distinction. Large glomeruloid structures were not necessarily considered equivalent to a cribriform pattern as long as a single attachment or a space was present, as defined above. Intraglandular complexity was required for the definition of glomeruloid glands. Telescoping was excluded from this definition. Poorly formed glands were defined as ill-defined glands with poorly formed glandular lumina.23 IDCP was defined as malignant epithelial cells filling large acini and prostatic ducts with preservation of basal cells and having either: (a) a solid or dense cribriform pattern or (b) a loose cribriform or micropapillary pattern with either marked nuclear atypia (i.e. nuclear size 6 × normal or larger) or comedonecrosis.24,25 Prostatectomies were submitted to include the entire posterior gland and at least 75% of the total tissue. Any tumour nodules identified grossly were submitted in their entirety. Only cases with at least 1 year of clinical follow-up at Vanderbilt University Medical Center were included. Criteria of exclusion included radiation or androgen deprivation therapy prior to surgery, the absence of Gleason pattern 4 prostate cancer (grade group 1) and tumour volume < 5% on radical prostatectomy. Patients who underwent hormonal or radiation therapy prior to the prostate-specific antigen (PSA) recurrence threshold of 0.2 ng/ml were excluded from the analyses involving BCR, defined as post-prostatectomy PSA recurrence of 0.2 ng/ml (n = 4). All haematoxylin and eosin (H&E)-stained slides were re-reviewed and graded according to the 2014 modified Gleason system and immunohistochemistry scored by a genitourinary pathologist (G.A.G.).26 Grading and scoring reflected the dominant tumour. Clinical and pathological features evaluated included age, race, length of clinical follow-up, pathological stage, Gleason pattern, extraprostatic extension, lymph node status and BCR.
One representative tumour block per case containing the best representation of Gleason patterns 4 and IDCP present in the case was selected for immunohistochemistry. In some cases, a Gleason pattern 3 component of the dominant nodule may have not been present on the slide that was specifically stained, yet it was included in the description because it was part of the nodule. Gleason pattern 3 tumours away from the dominant tumour were not included in this study. Analysis was performed on individual primary and secondary patterns within a dominant tumour nodule [e.g. Gleason pattern 3 tumour part of a dominant 3 + 4 (grade group 2), 4 + 3 (grade group 3) or higher grade group tumours with a minor grade 3 component]. Thus, in this study we will refer to these individual tumour components as Gleason pattern 3 or Gleason pattern 4, not as grade groups 1 or 2. At the same time, we will use the grade group definition when referring to the entire tumour nodule.
IMMUNOHISTOCHEMISTRY
Immunohistochemistry for PTEN was performed according to the manufacturer’s instructions using a Leica-Bond Max immunostainer on 5-μm sections (PTEN, M3627; Dako, Carpinteria, CA, USA; 1:200, p63, GTX102425; Genetex, Irvine, CA, 1:100). We have previously established concordance of our in-house PTEN immunohistochemical assay with fluorescence in-situ hybridisation (FISH) status.27,28 IDCP with cribriform morphology was delineated from invasive cribriform prostate cancer by p63 immunohistochemistry to assess for presence of basal cells, and diagnosis of IDCP was rendered based on established criteria.24 PTEN cytoplasmic and nuclear staining in ≥ 10% of cells was defined as positive. Homogenous PTEN loss was defined by a negative stain (≤ 10%) or markedly decreased expression compared to surrounding benign tissue and/or stroma in > 10% of tumour cells. Heterogenous PTEN loss was defined by focal loss of PTEN in > 10% and < 100% of the tumour cells, either within the dominant nodule or individual glands.18
STATISTICAL ANALYSIS
All analyses were conducted in R version 4.0.2. For all analyses, statistical significance was considered to be achieved at the nominal α = 0.05 level (two-sided). We computed descriptive statistics as median and interquartile range (IQR) for continuous variables or absolute and relative frequencies for categorical variables, as appropriate. Missing data were addressed using multiple imputations with chained equations,29 based on median = 500 iterations. Results were aggregated using Rubin’s rules.30
To evaluate predictors of cancer stage we used logistic regression with stage as the outcome (T3a or T3b versus T2), including the following as predictors of interest: age, pre-operative PSA, Gleason grade 3, Gleason grade 5, glomeruloid pattern, invasive cribriform carcinoma, poorly formed pattern and IDCP.
We further sought to evaluate predictors of BCR. To account for death as a competing risk for BCR, we employed a modification to the traditional Cox model based on the method described by Fine & Gray, in which the competing event of death is no longer treated as a censoring event.31 In accordance with their methods, we refer to the hazard ratios (HR) as ‘subdistribution hazard ratios’ throughout this work, although they can be conceptualised analogously to traditional HRs. In this modified Cox proportional hazards model, we included all the predictors of the previous model and the following additional predictors: number of lymph nodes removed, number of positive lymph nodes, positive margins and categorical cancer stage. We then followed-up with an analogous exploratory analysis in which invasive cribriform carcinoma and IDCP were considered in a composite fashion.
To evaluate and compare the odds of PTEN loss within the specific morphology patterns, we used generalised estimating equations with a working independence correlation structure and a logistic link function.32 We conducted an additional analysis of time to BCR that included indicators for PTEN loss within each of the following morphologic groups: Gleason grade 3, IDCP, glomeruloid pattern, invasive cribriform carcinoma and poorly formed pattern.
Results
We identified 163 prostatectomy cases that met inclusion criteria (Table 1). The median age was 63 years (IQR = 59–68) and the median pre-operative PSA was 7.95 ng/ml (IQR = 5.45–12.1). Our cohort was predominantly white (93%) and had a median clinical follow-up of 60 (IQR = 29.5–83.5) months. BCR occurred in 76 patients and death occurred in 7 patients, 1 of which was due to prostate cancer. Grade group distribution was as follows: grade group 2, 29 of 163 (17.8%); grade group 3, 42 of 163 (25.8%); grade group 4, 45 of 163 (27.6%); and grade group 5, 47 of 163 (28.8%). Most patients (70.6%) underwent pelvic lymph node dissection and 15.7% had nodal metastases. Any PTEN loss (homogenous or heterogeneous) was seen in 55 of 95 (58%) of Gleason pattern 3, 72 of 112 (64%) of cribriform, 87 of 141 (62%) of poorly formed, 18 of 49 (37%) of glomeruloid patterns and 30 of 35 (86%) of IDCP (Figures 1 and 2; Table 2).
Table 1.
Clinical and pathological characteristics
| Variable | Median [IQR] or n (%) |
|---|---|
| Pre-operative PSA (ng/ml); median [IQR] | 7.95 [5.45, 12.1] |
| Age (years); median [IQR] | 63 [59, 68] |
| Non-white race | 11/163 (6.75) |
| Grade group | |
| 2 | 29/163 (17.8) |
| 3 | 42/163 (25.8) |
| 4 | 45/163 (27.6) |
| 5 | 47/163 (28.8) |
| Positive margins | 50/163 (30.7) |
| Stage | |
| 2 | 70/163 (42.9) |
| 3a | 69/163 (42.3) |
| 3b | 24/163 (14.7) |
| Lymph node dissection | 115/163 (70.6) |
| Positive lymph nodes | 18/115 (15.7) |
| Biochemical recurrence | 76 (47) |
| Overall death | 7 (4) |
| Follow-up (months); median [IQR] | 60 [29.5, 83.5] |
| Time to BCR (months); median [IQR] | 22 [7.75, 36.8] |
BCR, biochemical recurrence; IQR, interquartile range; PSA, prostate-specific antigen (excludes patients who never reached a PSA nadir)
Figure 1.
Phosphatase and tensin homologue (PTEN) expression in poorly formed and glomeruloid Gleason pattern 4 carcinoma. Poorly formed glands (A,C) with PTEN loss (B) and intact PTEN (D). Glomeruloid pattern (E,G) with PTEN loss (F) and intact PTEN (H).
Figure 2.
Phosphatase and tensin homologue (PTEN) expression in invasive cribriform and intraductal carcinoma (IDCP). Invasive cribriform carcinoma (A,D) with PTEN loss (C) and intact PTEN (F). IDCP (G,J) with PTEN loss (I) and intact PTEN (L). The central panel shows basal cell loss (B,E) and basal cell preservation (H,K) by 63 immunohistochemistry.
Table 2.
PTEN expression by prostate cancer Gleason pattern/morphology
| Grade | n (%) |
|---|---|
| Gleason 3, n = 132 | |
| No loss | 40/95 (42.1) |
| Heterogeneous loss | 23/95 (24.2) |
| Homogenous loss | 32/95 (33.7) |
| Not performed | 37/132 (28.0) |
| Gleason 4 cribriform, n = 113 | |
| No loss | 40/112 (35.7) |
| Heterogeneous loss | 31/112 (27.7) |
| Homogenous loss | 41/112 (36.6) |
| Not performed | 1/113 (0.88) |
| Gleason 4 poorly formed, n = 141 | |
| No loss | 54/141 (38.3) |
| Heterogeneous loss | 31/141 (22.0) |
| Homogenous loss | 56/141 (39.7) |
| Not performed | 0/141 (0.0) |
| Gleason 4 glomeruloid, n = 51 | |
| No loss | 31/49 (63.2) |
| Heterogeneous loss | 8/49 (16.3) |
| Homogenous loss | 10/49 (20.4) |
| Not performed | 2/51 (3.92) |
| Intraductal carcinoma, n = 38 | |
| No loss | 5/35 (14.3) |
| Heterogeneous loss | 6/35 (17.1) |
| Homogenous loss | 24/35 (68.6) |
| Not performed | 3/38 (7.89) |
PTEN, phosphatase and tensin homologue.
Patients who had Gleason pattern 3 [i.e. compared to absence of Gleason pattern 3; OR = 0.23 (0.069, 0.75), P = 0.015] or glomeruloid pattern [OR = 0.39 (0.17, 0.93), P = 0.033] had lower odds of stage pT3a or pT3b. Additionally, Gleason pattern 5 [OR = 3.44 (1.60, 7.390, P = 0.002] and pre-operative PSA [OR = 1.06 (1.01, 1.12), P = 0.029] were associated with higher odds of stage pT3a or pT3b (Supporting information, Table S1).
The presence of Gleason pattern 3 was associated with a lower cumulative incidence of BCR by the subdistribution hazard model [HR = 0.48 (0.28, 0.82), P = 0.006]. The presence of invasive cribriform carcinoma was associated with a higher cumulative incidence of BCR [HR = 3.50 (1.89, 6.49), P < 0.001]; our data did not find sufficient evidence of an association between IDCP and cumulative incidence of BCR [HR = 1.07 (0.62, 1.87), P = 0.81] (Table 3). In the exploratory model in which invasive cribriform carcinoma and IDCP were considered together as a composite (Table 4), we found evidence of an association with cumulative incidence of BCR [HR = 5.06 (2.21, 11.6), P < 0.001] (Figure 3).
Table 3.
Multivariable competing risks regression model using subdistribution hazard ratios to compare the cumulative incidence of biochemical recurrence across morphology indicators
| Variable | SHR (95% CI) | P |
|---|---|---|
| Gleason 3 | 0.48 (0.28, 0.82) | 0.006 |
| Gleason 5 | 1.11 (0.65, 1.89) | 0.72 |
| Intraductal carcinoma | 1.07 (0.62, 1.87) | 0.81 |
| Gleason 4 subpattern | ||
| Glomeruloid | 0.58 (0.31, 1.08) | 0.086 |
| Cribriform | 3.50 (1.89, 6.49) | < 0.001 |
| Poorly formed | 1.11 (0.54, 2.29) | 0.78 |
| Age (years) | 0.998 (0.96, 1.03) | 0.93 |
| Pre-operative PSA (units) | 1.003 (1.000, 1.006) | 0.052 |
| Lymph nodes removed | 2.14 (1.01, 4.50) | 0.046 |
| Lymph nodes positive | 1.86 (0.95, 3.64) | 0.072 |
| Positive margins | 1.17 (0.69, 1.97) | 0.56 |
| Stage T2** | Ref. | Ref. |
| Stage T3a | 2.92 (1.49, 5.72) | 0.002 |
| Stage T3b | 5.99 (2.18, 16.5) | 0.001 |
CI, confidence interval; SHR, subdistribution hazard ratio.
Statistically significant P values are highlighted in bold.
Omnibus test of association between stage and cumulative incidence of biochemical recurrence: P < 0.001.
Table 4.
Multivariable competing risks regression model using subdistribution hazard ratios to compare the cumulative incidence of biochemical recurrence across morphology indicators combining invasive cribriform cancer and intraductal carcinoma
| Variable | SHR (95% CI) | P |
|---|---|---|
| Gleason 3 | 0.52 (0.31, 0.89) | 0.016 |
| Gleason 5 | 1.12 (0.64, 1.95) | 0.70 |
| Gleason 4 subpattern | ||
| Glomeruloid | 0.74 (0.39, 1.41) | 0.36 |
| Cribriform or intraductal carcinoma | 5.06 (2.21, 11.6) | < 0.001 |
| Poorly formed | 1.01 (0.49, 2.10) | 0.97 |
| Age (years) | 0.99 (0.96, 1.03) | 0.67 |
| Pre-operative PSA (units) | 1.003 (1.000, 1.006) | 0.029 |
| Lymph nodes removed | 1.94 (0.88, 4.26) | 0.10 |
| Lymph nodes positive | 1.78 (0.93, 3.42) | 0.083 |
| Positive margins | 1.32 (0.78, 2.23) | 0.30 |
| Stage T2** | Ref. | Ref. |
| Stage T3a | 2.78 (1.43, 5.43) | 0.003 |
| Stage T3b | 4.47 (1.61, 12.5) | 0.004 |
CI, confidence interval; PSA, prostate-specific antigen; SHR, subdistribution hazard ratio.
Statistically significant P values are highlighted in bold.
Omnibus test of association between stage and cumulative incidence of biochemical recurrence: P < 0.001.
Figure 3.
Cumulative incidence of biochemical recurrence (BCR) by presence of any cribriform pattern (invasive cribriform carcinoma or intraductal carcinoma).
Compared to Gleason pattern 3, glomeruloid morphology was associated with lower odds of PTEN loss [OR = 0.45 (0.24, 0.82), P = 0.009], while IDCP had higher odds of PTEN loss [OR = 4.232 (1.60, 11.20, P = 0.004] (Supporting information, Table S2). When analysing the cumulative incidence of BCR regardless of the presence of a specific morphology and accounting for death as a competing risk, invasive cribriform carcinoma remained predictive of BCR [OR = 3.72 (1.75, 7.94), P = 0.001], but not PTEN loss within invasive cribriform carcinoma. Interestingly, glomeruloid morphology was associated with a lower hazard of BCR [OR = 0.27 (0.088, 0.80), P = 0.018], but PTEN loss within glomeruloid morphology was associated with a higher hazard of BCR [OR = 4.07 (1.04, 15.9), P = 0.043] (Table 5). There was no sufficient evidence of an association between PTEN expression in Gleason pattern 3 or glomeruloid pattern and adjacent pattern 4 morphologies or IDCP (Supporting information, Tables S3 and S4).
Table 5.
Multivariable competing risks regression model using subdistribution hazard ratios to compare the cumulative incidence of biochemical recurrence across morphology and respective PTEN loss indicators
| Variable | SHR (95% CI) | P |
|---|---|---|
| Gleason 3 | 0.75 (0.40, 1.39) | 0.35 |
| Gleason 5 | 1.004 (0.99, 1.02) | 0.62 |
| Gleason 4 subpattern | ||
| Glomeruloid | 0.27 (0.088, 0.80) | 0.018 |
| Cribriform | 3.72 (1.75, 7.94) | 0.001 |
| Poorly formed | 1.81 (0.69, 4.80) | 0.23 |
| Intraductal carcinoma | 1.10 (0.30, 4.04) | 0.89 |
| Age | 0.999 (0.96, 1.04) | 0.95 |
| Pre-operative PSA | 1.002 (0.999, 1.006) | 0.18 |
| Lymph nodes removed | 1.97 (0.96, 4.02) | 0.063 |
| Lymph nodes positive | 1.95 (0.88, 4.34) | 0.10 |
| Positive margins | 1.04 (0.57, 1.91) | 0.89 |
| Stage T2** | Ref. | Ref. |
| Stage T3a | 2.99 (1.43, 6.25) | 0.004 |
| Stage T3b | 7.32 (2.53, 21.2) | < 0.001 |
| PTEN loss | ||
| Gleason 3 | 0.54 (0.25, 1.16) | 0.11 |
| Gleason 4 glomeruloid | 4.07 (1.04, 15.9) | 0.043 |
| Gleason 4 cribriform | 1.10 (0.50, 2.41) | 0.81 |
| Gleason 4 poorly formed | 0.57 (0.21, 1.53) | 0.26 |
| Intraductal carcinoma | 0.98 (0.24, 4.11) | 0.98 |
CI, confidence interval; PTEN, phosphatase and tensin homologue; SHR, subdistribution hazard ratio.
Statistically significant P values are highlighted in bold.
Omnibus test of association between stage and cumulative incidence of biochemical recurrence: P < 0.001.
Discussion
Although it has long been recognised that PTEN loss is a common occurrence in prostate cancer, relatively few studies have evaluated the association of PTEN loss with morphological features in prostate cancer, mainly separately evaluating cancer morphology and clinical outcomes. Previous studies have shown that PTEN is frequently lost in IDCP.21 Similarly, we found PTEN loss in a high percentage of IDCP (79%). When comparing different patterns, PTEN loss was significantly more frequent in IDCP [OR = 4.232 (1.596, 11.22), P = 0.004]. Other studies have evaluated PTEN loss among different prostate cancer patterns.19,33 In a study by Ronen et al., PTEN loss was significantly associated with cribriform pattern in 52 cases, of which 46 had both cribriform and benign pattern for comparison.19 However, glomeruloid pattern or IDCP were not assessed in this study. Similarly, PTEN loss was significantly associated with IDCP and invasive cribriform carcinoma in a study from 260 prostate biopsies.33 Our study evaluates the association of both PTEN loss and cancer morphology within the same cohort and attempts to address the impact of Gleason pattern 4 and PTEN expression on clinical outcomes, evaluated as competing independent variables. As expected, we found that increasing Gleason pattern was predictive of stage pT3, with Gleason pattern 5 being significantly associated with non-organ-confined disease. However, we did not find sufficient evidence of an association between either invasive cribriform carcinoma or IDCP and cancer stage, although the presence of any cribriform morphology (IDCP or invasive cribriform carcinoma) was associated with a higher cumulative incidence of BCR. When analysed individually, only invasive cribriform carcinoma remained a significant predictor of BCR. Thus, in our study, the excess hazard associated with the composite of IDCP and invasive cribriform carcinoma is primarily driven by the invasive cribriform carcinoma component rather than by IDCP. As such, it may be beneficial to analyse invasive cribriform carcinoma and IDCP separately in future studies evaluating cribriform morphology and clinical outcomes. Although invasive cribriform carcinoma was a significant predictor of BCR, and IDCP had the highest relative risk for PTEN loss, neither PTEN loss within invasive cribriform carcinoma nor PTEN loss within IDCP predicted BCR. Thus, invasive cribriform carcinoma was associated with a higher hazard of BCR, independently of the effect of PTEN loss. These data do not sufficiently support the hypothesis that further PTEN loss within an invasive cribriform tumour is associated with an additional hazard of BCR beyond that of invasive cribriform carcinoma alone.
Gleason pattern 3 had a high incidence of PTEN loss in our study (58%). This finding is in apparent conflict with a previously reported incidence of 5% (2–7%) for grade group 1 in biopsy cohorts. However, the incidence of PTEN loss increases to 18–27% in biopsies from grade group 1 tumours that are upgraded in subsequent biopsies or radical prostatectomies.28,34–36 Our PTEN analysis was conducted in Gleason pattern 3 components of high-grade dominant tumours, 82% of which were grade groups 3–5. This PTEN loss difference between a Gleason pattern 3 component of grade group 1 and grade groups 2–5 tumours support prior observations of intra- and intertumoral heterogeneity and, conversely, the notion that adjacent Gleason patterns 3 and 4 tumours could be clonally linked, or that a subset of Gleason pattern 3 tumours progresses to 4 or derives from a common precursor.37–39
We report PTEN loss to be significantly less common in glomeruloid pattern. In concordance with this finding, we found glomeruloid pattern to be associated with organ-confined disease and a decreased risk of BCR. Other studies have also reported better clinical outcomes in glomeruloid than cribriform pattern, although worse than grade group 1 cancer.9,40 In our study, PTEN loss was associated with biochemical recurrence within the glomeruloid pattern but not in other patterns. It is possible, although purely speculative, that glomeruloid pattern is a transitional pattern between Gleason patterns 3 and 4 cribriform carcinoma and, thus still a harbinger of PTEN-dependent outcome. Clonal evolution into cribriform morphology, independent of PTEN loss, could represent the molecular underpinning of the poor prognostic effect of invasive cribriform carcinoma of the prostate.41
Our findings regarding the effects of glomeruloid pattern are novel. As seen in other studies, glomeruloid pattern was associated with a lower hazard of BCR. However, our data provide evidence that PTEN loss within glomeruloid pattern is associated with an increased risk of BCR compared to glomeruloid pattern without PTEN loss. These findings may have potential clinical utility. It has been previously suggested that PTEN could potentially be used as a biomarker to help stratify patients at risk for disease progression. It is often the case that patients with low- to intermediate-risk disease are placed on active surveillance. In the correct clinical context, a patient with low-volume grade group 2 disease with a low percentage of pattern 4 could be a candidate for active surveillance. Studies have suggested that cribriform pattern on prostate biopsy be a contraindication for active surveillance.2 Similarly, a patient with intermediate-risk disease and glomeruloid pattern on prostate biopsy may benefit from PTEN testing to direct management. However, these findings need validation in larger studies.
Our study has several limitations, including its retrospective nature and exploration of patients with higher grade/stage cancers. Indeed, none of our cases were grade group 1 on radical prostatectomy, per study design. This is evidenced by more than 55% of our cohort having non-organ-confined disease and more than 40% of patients experiencing BCR. As such, these results need to be interpreted within this high-risk population. In addition, there was only partial submission of prostate tissue. However, we evaluated the entire posterior gland, all suspicious lesions and at least 75% of each specimen, and are thereby confident that our tissue submission was adequate for the appropriate staging and grading of specimens. We acknowledge that many variables were tested in the multivariable competing risks regression analysis of BCR across morphology and respective PTEN loss indicators, that multiple testing can inflate type 1 error rates and that the number of events was low for the number of variables included in our multivariable analyses; however, despite possible concerns of overfitting, we were able to obtain statistically significant evidence of a number of adjusted associations. Further confirmatory studies will need to be conducted to substantiate these analyses, which were exploratory in nature. Furthermore, although our cohort is one of the largest to date to examine PTEN expression within different architectural patterns, the number of cases investigated is still relatively low, especially as it applies to the number of IDCP cases. Despite these limitations, this is the first study, to our knowledge, to evaluate both tumour morphology and PTEN expression as competing variables in predicting clinical outcomes and to separately distinguish the effect of invasive cribriform carcinoma from IDCP.
Conclusions
Gleason pattern 3 and glomeruloid pattern were associated with lower odds of stage pT3 prostate cancer and BCR. However, within the glomeruloid pattern, PTEN loss was a significant predictor of BCR. The presence of any cribriform pattern was associated with BCR, although this finding was primarily driven by the invasive cribriform tumour in our study. This effect was maintained despite a lack of association between PTEN loss and invasive cribriform carcinoma. We speculate that other molecular drivers independent from PTEN loss could contribute to the poor prognostic effect of invasive cribriform carcinoma of the prostate.
Supplementary Material
Table S1. Logistic regression model comparing the odds of cancer stage pT3 relative to stage pT2 across Gleason patterns and morphologic patterns.
Table S2. Generalized estimating equations to compare the odds of PTEN loss (homozygous or heterozygous) across different patterns.
Table S3. Predictors of PTEN loss within Gleason pattern 3.
Table S4. Predictors of PTEN loss within glomeruloid pattern.
Acknowledgements
We thank Kelli Boyd MD, Cindy Lowe BS, HTL (ASCP), QIHC, Sherry Smith, Connie Nixon and the Vanderbilt Translational Pathology Shared Resource for their assistance with the histological and immunohistochemical portion of the study.
Footnotes
Conflicts of interest
All authors report no conflicts of interest or financial disclosures that were pertinent to the following study.
Supporting Information
Additional Supporting Information may be found in the online version of this article:
References
- 1.Siegel RL, Miller KD, Fuchs HE, Jemal A. Cancer Statistics, 2021. CA: Cancer J. Clin 2021; 71; 7–33. 10.3322/caac.21654 [DOI] [PubMed] [Google Scholar]
- 2.Epstein JI, Amin MB, Fine SW et al. The 2019 Genitourinary Pathology Society (GUPS) White Paper on contemporary grading of prostate cancer. Arch. Pathol. Lab. Med 2021; 145; 461–493. [DOI] [PubMed] [Google Scholar]
- 3.Offermann A, Hupe MC, Sailer V, Merseburger AS, Perner S. The new ISUP 2014/WHO 2016 prostate cancer grade group system: first resume 5 years after introduction and systemic review of the literature. World J. Urol 2020; 38; 657–662. [DOI] [PubMed] [Google Scholar]
- 4.Paner GP, Stadler WM, Hansel DE, Montironi R, Lin DW, Amin MB. Updates in the eighth edition of the tumour–node–metastasis staging classification for urologic cancers. Eur. Urol 2018; 73; 560–569. [DOI] [PubMed] [Google Scholar]
- 5.Downes MR, Xu B, van der Kwast TH. Cribriform architecture prostatic adenocarcinoma in needle biopsies is a strong independent predictor for lymph node metastases in radical prostatectomy. Eur. J. Cancer 2021; 148; 432–439. [DOI] [PubMed] [Google Scholar]
- 6.Keefe DT, Schieda N, El Hallani S et al. Cribriform morphology predicts upstaging after radical prostatectomy in patients with Gleason score 3 + 4 = 7 prostate cancer at transrectal ultra-sound (TRUS)-guided needle biopsy. Virchows Arch 2015; 467; 437–442. [DOI] [PubMed] [Google Scholar]
- 7.Kweldam CF, Kummerlin IP, Nieboer D et al. Presence of invasive cribriform or intraductal growth at biopsy outperforms percentage grade 4 in predicting outcome of Gleason score 3+4 = 7 prostate cancer. Mod. Pathol 2017; 30; 1126–1132. [DOI] [PubMed] [Google Scholar]
- 8.Montironi R, Zhou M, Magi-Galluzzi C, Epstein JI. Features and prognostic significance of intraductal carcinoma of the prostate. Eur. Urol. Oncol 2018; 1; 21–28. [DOI] [PubMed] [Google Scholar]
- 9.Choy B, Pearce SM, Anderson BB et al. Prognostic significance of percentage and architectural types of contemporary Gleason pattern 4 prostate cancer in radical prostatectomy. Am. J. Surg. Pathol 2016; 40; 1400–1406. [DOI] [PubMed] [Google Scholar]
- 10.Dong F, Yang P, Wang C et al. Architectural heterogeneity and cribriform pattern predict adverse clinical outcome for Gleason grade 4 prostatic adenocarcinoma. Am. J. Surg. Pathol 2013; 37; 1855–1861. [DOI] [PubMed] [Google Scholar]
- 11.Kweldam CF, Wildhagen MF, Steyerberg EW et al. Cribriform growth is highly predictive for postoperative metastasis and disease-specific death in Gleason score 7 prostate cancer. Mod. Pathol 2015; 28; 457–464. [DOI] [PubMed] [Google Scholar]
- 12.McKenney JK, Wei W, Hawley S et al. Histologic grading of prostatic adenocarcinoma can be further optimized: analysis of the relative prognostic strength of individual architectural patterns in 1275 patients from the Canary Retrospective Cohort. Am. J. Surg. Pathol 2016; 40; 1439–1456. [DOI] [PubMed] [Google Scholar]
- 13.Chua MLK, Lo W, Pintilie M et al. A prostate cancer ‘nimbosus’: genomic instability and SChLAP1 dysregulation underpin aggression of intraductal and cribriform subpathologies. Eur. Urol 2017; 72; 665–674. [DOI] [PubMed] [Google Scholar]
- 14.Krohn A, Freudenthaler F, Harasimowicz S et al. Heterogeneity and chronology of PTEN deletion and ERG fusion in prostate cancer. Mod. Pathol 2014; 27; 1612–1620. [DOI] [PubMed] [Google Scholar]
- 15.Ullman D, Dorn D, Rais-Bahrami S, Gordetsky J. Clinical utility and biologic implications of phosphatase and tensin homolog (PTEN) and ETS-related gene (ERG) in prostate cancer. Urology 2018; 113; 59–70. [DOI] [PubMed] [Google Scholar]
- 16.Krohn A, Diedler T, Burkhardt L et al. Genomic deletion of PTEN is associated with tumour progression and early PSA recurrence in ERG fusion-positive and fusion-negative prostate cancer. Am. J. Pathol 2012; 181; 401–412. [DOI] [PubMed] [Google Scholar]
- 17.Jamaspishvili T, Berman DM, Ross AE et al. Clinical implications of PTEN loss in prostate cancer. Nat. Rev. Urol 2018; 15; 222–234. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Lotan TL, Heumann A, Rico SD et al. PTEN loss detection in prostate cancer: comparison of PTEN immunohistochemistry and PTEN FISH in a large retrospective prostatectomy cohort. Oncotarget 2017; 8; 65566–65576. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Ronen S, Abbott DW, Kravtsov O et al. PTEN loss and p27 loss differ among morphologic patterns of prostate cancer, including cribriform. Hum. Pathol 2017; 65; 85–91. [DOI] [PubMed] [Google Scholar]
- 20.Downes MR, Satturwar S, Trudel D, van der Kwast TH. Evaluation of ERG and PTEN protein expression in cribriform architecture prostate carcinomas. Pathol. Res. Pract 2017; 213; 34–38. [DOI] [PubMed] [Google Scholar]
- 21.Morais CL, Han JS, Gordetsky J et al. Utility of PTEN and ERG immunostaining for distinguishing high-grade PIN from intraductal carcinoma of the prostate on needle biopsy. Am. J. Surg. Pathol 2015; 39; 169–178. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Pacelli A, Lopez-Beltran A, Egan AJ, Bostwick DG. Prostatic adenocarcinoma with glomeruloid features. Hum. Pathol 1998; 29; 543–546. [DOI] [PubMed] [Google Scholar]
- 23.Epstein JI, Allsbrook WC Jr, Amin MB, Egevad LL, ISUP Grading Committee. The 2005 International Society of Urological Pathology (ISUP) Consensus Conference on Gleason grading of prostatic carcinoma. Am. J. Surg. Pathol 2005; 29(9); 1228–1242. [DOI] [PubMed] [Google Scholar]
- 24.Guo CC, Epstein JI. Intraductal carcinoma of the prostate on needle biopsy: histologic features and clinical significance. Mod. Pathol 2006; 19; 1528–1535. [DOI] [PubMed] [Google Scholar]
- 25.Moch HHP, Ulbright TM, Reuter VE. WHO Classification of tumours of the urinary system and male genital organs Lyon, France: International Agency for Research on Cancer; 2016. [Google Scholar]
- 26.Epstein JI, Egevad L, Amin MB et al. The 2014 International Society of Urological Pathology (ISUP) Consensus Conference on Gleason grading of prostatic carcinoma: definition of grading patterns and proposal for a new grading system. Am. J. Surg. Pathol 2016; 40; 244–252. [DOI] [PubMed] [Google Scholar]
- 27.Goldstein J, Borowsky AD, Goyal R et al. MAGI-2 in prostate cancer: an immunohistochemical study. Hum. Pathol 2016; 52; 83–91. [DOI] [PubMed] [Google Scholar]
- 28.Picanço-Albuquerque CG, Morais CL, Carvalho FLF et al. In prostate cancer needle biopsies, detections of PTEN loss by fluorescence in situ hybridization (FISH) and by immunohistochemistry (IHC) are concordant and show consistent association with upgrading. Virchows Arch 2016; 468; 607–617. [DOI] [PubMed] [Google Scholar]
- 29.van Buuren S Multiple imputation of discrete and continuous data by fully conditional specification. Stat. Methods Med. Res 2007; 16; 219–242. [DOI] [PubMed] [Google Scholar]
- 30.Rubin DB. Multiple imputation for nonresponse in surveys New York, NY: John Wiley & Sons; 1987. [Google Scholar]
- 31.Fine J, Gray RJ. Proportional hazards model for the subdistribution of a competing risk. J. Am. Stat. Assoc 1999; 94; 496–509. [Google Scholar]
- 32.Diggle PJ, Heagerty P, Liang K, Zeger SL. Analysis of longitudinal data 2nd edn. Oxford: Oxford University Press; 2014. [Google Scholar]
- 33.Shah RB, Shore KT, Yoon J, Mendrinos S, McKenney JK, Tian W. PTEN loss in prostatic adenocarcinoma correlates with specific adverse histologic features (intraductal carcinoma, cribriform Gleason pattern 4 and stromogenic carcinoma). Prostate 2019; 79; 1267–1273. [DOI] [PubMed] [Google Scholar]
- 34.Tosoian JJ, Guedes LB, Morais CL et al. PTEN status assessment in the Johns Hopkins active surveillance cohort. Prostate Cancer Prostat. Dis 2019; 22; 176–181. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Lotan TL, Carvalho FLF, Peskoe SB et al. PTEN loss is associated with upgrading of prostate cancer from biopsy to radical prostatectomy. Mod. Pathol 2015; 28; 128–137. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Trock BJ, Fedor H, Gurel B et al. PTEN loss and chromosome 8 alterations in Gleason grade 3 prostate cancer cores predicts the presence of un-sampled grade 4 tumour: implications for active surveillance. Mod. Pathol 2016; 29; 764–771. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Kovtun IV, Cheville JC, Murphy SJ et al. Lineage relationship of Gleason patterns in Gleason score 7 prostate cancer. Cancer Res 2013; 73; 3275–3284. [DOI] [PubMed] [Google Scholar]
- 38.Sowalsky AG, Ye H, Bubley GJ, Balk SP. Clonal progression of prostate cancers from Gleason grade 3 to grade 4. Cancer Res 2013; 73; 1050–1055. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Sowalsky AG, Kissick HT, Gerrin SJ et al. Gleason score 7 prostate cancers emerge through branched evolution of clonal Gleason pattern 3 and 4. Clin. Cancer Res 2017; 23; 3823–3833. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Hollemans E, Verhoef EI, Bangma CH et al. Clinicopathological characteristics of glomeruloid architecture in prostate cancer. Mod. Pathol 2020; 33; 1618–1625. [DOI] [PubMed] [Google Scholar]
- 41.Böttcher R, Kweldam CF, Livingstone J et al. Cribriform and intraductal prostate cancer are associated with increased genomic instability and distinct genomic alterations. BMC Cancer 2018; 18; 8. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1. Logistic regression model comparing the odds of cancer stage pT3 relative to stage pT2 across Gleason patterns and morphologic patterns.
Table S2. Generalized estimating equations to compare the odds of PTEN loss (homozygous or heterozygous) across different patterns.
Table S3. Predictors of PTEN loss within Gleason pattern 3.
Table S4. Predictors of PTEN loss within glomeruloid pattern.