The importance of the season of biopsy on the Gleason score on biopsy: does exposure to sunshine have an influence?

Abstract

The circadian clock is strongly influenced by the sun exposure and prostate cancer has been shown to be inversely proportional to it. We investigated whether PCa aggressiveness in Montreal, Quebec, Canada, differs over the months during or following potentially longer exposure to sunlight. We analyzed 3447 patients treated between January 1995 and December 2023 with primary radiotherapy for localized PCa. We investigated whether the month when diagnostic biopsy was performed was associated with a more frequent diagnosis of a primary Gleason score (pGS) of 4 or 5. We grouped the months of biopsy into the four quarters (Q1–4) of the year. Multivariable logistic regression was used to predict a pGS of 4 or 5, adjusted for age and year of biopsy. There were significantly fewer biopsies (P = 0.027) with pGS 4 or 5 in the last 3 months of the year (Q4; 19.0%) than those in Q1–3 (22.9%). Age, prostate-specific antigen (PSA) level, and the number of positive biopsies were not significantly different between Q4 versus Q1–3. In multivariate logistic regression analysis, a biopsy in Q4 was significantly predictive of a lower risk of pGS 4 or 5 (odds ratio [OR]: 0.77, 95% confidence interval [CI]: 0.63–0.93, P = 0.007), as was older age (P < 0.001), but not the year of biopsy (P = 0.76). In conclusion, patients biopsied during Q4 had a 23% lower risk of a pGS 4 or 5 on diagnostic biopsy than those biopsied during the previous 9 months. Our results are not a proof of causality.

INTRODUCTION

The influence of sunlight exposure on prostate cancer (PCa) has long been known. A North–South trend for PCa mortality in the USA was reported in 1992, with lower rates in the South.1 This relationship between sun exposure and PCa is probably meditated by vitamin D through an effect on cell differentiation and proliferation.2 Large studies such as the First National Health and Nutrition Examination Survey (NHANES) have shown that residence in the South of the USA with higher exposure to solar radiation were associated with a significant reduction in PCa risk.3

The region of residence does not seem to be the only factor influencing exposure to sunlight and PCa. Shift work also appears to have an impact on PCa. NHANES data showed that shift work was associated with higher prostate-specific antigen (PSA) levels.4 A possible explanation could be that shift work results in decreased exposure to sunlight, which, in turn, can disrupt the circadian rhythm. The relationship between the circadian clock and PCa is supported by the fact that the circadian clock may function as a tumor suppressor and act on cell proliferation, apoptosis, cell cycle control, and DNA damage response.5

In Canada, winters are characterized by the prolonged periods of frigid temperatures. This can lead to a reduction in direct sun exposure, which might influence PCa aggressiveness. In this study, we examined our institutional database (Department of Radiation Oncology, University of Montreal Health Center, Montréal, Quebec, Canada) of patients treated with primary radiotherapy for PCa to determine whether the aggressiveness of the disease varies based on the month of diagnosis, specifically considering potential differences in sunlight exposure.

PARTICIPANTS AND METHODS

All included patients were registered in our institutional database. This database included patients treated with brachytherapy or external-beam radiotherapy (EBRT) for primary PCa. The study was approved by the institution’s ethics committee (Research Centre of University of Montreal Health Center; Approval No. CER 23.177). The need for signed informed consent was waived because of the retrospective nature of the study. Patients who underwent prostatectomy before EBRT were excluded from this study. For patients who had several biopsies before treatment, the last biopsy for this analysis was used. Of the 4532 patients treated with primary radiotherapy, 785 (17.3%) were excluded because they had incomplete data on date of biopsy or Gleason score, with 3747 patients remaining for data analysis.

The average monthly hours of sunshine in Montreal, Quebec, Canada depend greatly on the season: the highest number of hours of sunshine is in July with about 272 h, and the least sunny is December with about 80 h. The average monthly sunshine duration from May to August is >220 h. All other months are <180 h.6

Patients biopsied during the first three quarters (Q1–3, January to September) were compared to those biopsied in the last quarter (Q4, October to December) using the Chi-squared test for categorical variables (with continuity correction) and one-way analysis of variance (ANOVA) test for continuous variables, with equal variance assumption.

Logistic regression analysis was used with a primary Gleason score (pGS) of 4/5 as the dependent variable. The independent variable as a continuous variable was age, and as a binary factor was the date of biopsy (Q4 versus Q1–3) and whether the biopsy was performed ≤2005 or after 2005 because of a change in Gleason grading in 2004 such as for example a refinement of Gleason pattern 4.7 P < 0.05 was considered statistically significant.

RESULTS

Of the 3747 analyzed patients, 900 (24.0%) underwent biopsy in the last 3 months of the year (Q4, October to December). Patients’ characteristics are listed in Table 1. Biopsies performed in the last quarter of the year (Q4) were significantly less likely (P = 0.027) to have a pGS of 4 or 5 (19.0%) than those performed in the nine preceding months (Q1–3, 22.9%). Age (P = 0.22), PSA level (P = 0.85), and the incidence of comorbidities such as diabetes or dyslipidemia did not differ between Q4 and Q1–3. Patients in Q4 were more likely to have undergone a biopsy in or before 2005. Figure 1 illustrates the monthly distribution of pGS. To add more granularity, we analyzed each month individually. Only January reached borderline statistical significance (odds ratio [OR]: 0.64, 95% confidence interval [CI]: 0.42–0.99, P = 0.05).

Table 1.

Patient characteristics

Factor	Biopsy in January–September	Biopsy in October–December	P
Patient (n)	2847	900
Age (year), mean (s.d.)	66.7 (6.6)	67.0 (6.4)	0.222
PSA (ng ml−1), mean (s.d.)	8.9 (10.0)	8.98 (9.0)	0.847
Month of biopsy, n (%)			<0.001*
January	323 (11.3)
February	294 (10.3)
March	320 (11.2)
April	307 (10.8)
May	335 (11.8)
June	323 (11.3)
July	310 (10.9)
August	279 (9.8)
September	356 (12.5)
October		312 (34.7)
November		350 (38.9)
December		238 (26.4)
Primary Gleason score, n (%)			0.027*
Primary Gleason 4 or 5	651 (22.9)	171 (19.0)
Primary Gleason 3	2190 (76.9)	725 (80.6)
NA	6 (0.2)	4 (0.4)
CAPRA score, n (%)			0.009*
1	275 (9.7)	92 (10.2)
2	686 (24.1)	222 (24.7)
3	732 (25.7)	230 (25.6)
4	405 (14.2)	141 (15.7)
5	270 (9.5)	71 (7.9)
6	224 (7.9)	66 (7.3)
7	124 (4.4)	28 (3.1)
8	59 (2.1)	11 (1.2)
9	32 (1.1)	12 (1.3)
10	23 (0.8)	9 (1.0)
CAPRA score risk group, n (%)			0.001*
Low risk (score 0–2)	961 (33.8)	314 (34.9)
Intermediate risk (3–5)	1407 (49.4)	442 (49.1)
High risk (6–10)	462 (16.2)	126 (14.0)
NA	17 (0.6)	18 (2.0)
T-stage, n (%)			0.004*
T1 and T2	2687 (94.4)	843 (93.7)
T3 and T4	148 (5.2)	44 (4.9)
NA	12 (0.4)	13 (1.4)
Number of positive biopsies, n (%)			0.174
<34%	1337 (47.0)	434 (48.2)
≥34%	1509 (53.0)	464 (51.6)
NA	1 (<0.01)	2 (0.2)
Year of biopsy, n (%)			0.035*
≤2005	224 (7.9)	87 (9.7)	0.102
>2005	2623 (92.1)	813 (90.3)
PSA (ng ml−1), n (%)
≤6.0	1132 (39.8)	361 (40.1)	0.56
Between 6.1 and 10.0	1107 (38.9)	334 (37.1)
Between 10.1 and 20.0	459 (16.1)	150 (16.7)
Between 20.1 and 30.0	75 (2.6)	25 (2.8)
More than 30.0	68 (2.4)	25 (2.8)
NA	6 (0.2)	5 (0.6)
Treatment type, n (%)			0.603
Radiotherapy	977 (34.3)	289 (32.1)
Brachytherapy as monotherapy	1414 (49.7)	465 (51.7)
Brachytherapy boost	415 (14.6)	129 (14.3)
NA	41 (1.4)	17 (1.9)
Hormonal therapy, n (%)	478 (16.8)	140 (15.6)	0.413
Cardiovascular disease, n (%)	224 (7.9)	68 (7.6)	0.815
Treated for other disease, n (%)
Arrhythmia	112 (3.9)	32 (3.6)	0.678
Hypertension	1315 (46.2)	397 (44.1)	0.293
Diabetes	415 (14.6)	140 (15.6)	0.505
Dyslipidemia	1243 (43.7)	401 (44.6)	0.665

^*P<0.05. s.d.: standard deviation; PSA: prostate-specific antigen; CAPRA: cancer of the prostate risk assessment; NA: non-applicable

Figure 1.

Monthly distribution of patients with a primary Gleason score of 4/5 (blue) and a primary Gleason score of 3 (red).

In the multivariable analysis (Table 2), patients biopsied in Q4 were 23% less likely (95% CI: 7%–37%) to be diagnosed with a pGS of 4 or 5. Older age was also a significant predictor (OR: 1.09, 95% CI: 1.07–1.10, P < 0.001), but not if the biopsy was done before or after 2005 (P = 0.76). When adding the month of January to the Q4, therefore including the months October–January, the model lost its strong statistical significance (OR: 0.84, 95% CI: 0.71–1.0, P = 0.05).

Table 2.

Multivariable analysis to predict for primary Gleason 4 or 5

Factor	OR	95% CI	P
Biopsy in October–December vs January–September	0.77	0.63–0.93	0.007*
Age (year), continuous variable	1.09	1.07–1.10	<0.001*
Year of biopsy, ≤2005 vs >2005	1.04	0.79–1.38	0.76

*P<0.05. OR: odds ratio; CI: confidence interval

DISCUSSION

In this study, we found that patients treated with radiotherapy for PCa had 23% (95% CI: 7%–37%) less risk of having a pGS of 4 or 5 on diagnostic prostate biopsy when performed in the last quarter of the year (Q4) than during the other 9 months.

The reason why the biopsies in winter were more likely to show a pGS of 4 or 5 can be manifold. We believe that our data highlight the importance of environmental factors in PCa. The environment can influence vitamin levels, variations in hormone levels, healthcare access, and patient behavior, such as higher physical activity in summer and autumn. The fact that shift work has been shown to be associated with PCa risk points toward an environmental influence on PCa. The fact that shift work can cause cancer has been recognized.8 There are several reasons why shift work can lead to cancer. One is the disruption of circadian rhythms, which can suppress melatonin production, and lifestyle factors, such as irregular eating patterns and reduced physical activity and stress. A recently published meta-analysis8 showed an increase in PCa in shift workers, although a Canadian case–control study conducted in 2005–2012 in our province found no evidence of an increase in PCa in night shift workers.9

Cancer in general, and specifically, PCa, is reduced in men exposed to more sunlight. In a death-certification-based case–control study conducted in 2002, Freedman et al.10 found that residential exposure to sunlight was negatively and significantly associated with mortality from breast, ovarian, prostate, and colon cancer in women. However, mortality from ovarian and PCa is not consistently associated with occupational exposure to sunlight. Medium or high residential exposure to sunlight was associated with an approximately 10% reduction in PCa incidence compared with patients with low exposure. However, employment in farming jobs was associated with an increased risk of PCa. In a population-based, case–control study, sun exposure and polymorphisms in the vitamin D receptor (VDR) gene were found to be associated with the risk of primary advanced PCa, including cases classified as ≥T3, node-positive, or metastatic. There was a reduced risk associated with high sun exposure, high occupational outdoor activity, and putatively high-activity VDR genotypes in the presence of high sun exposure. The effect was important, with a risk reduction of 33%–54% in men with both high sun exposure and protective VDR genotypes.2 Unfortunately, we did not routinely measure vitamin D levels and were therefore unable to correlate the Gleason score with vitamin D levels.

The effects of sunlight exposure and season of diagnosis on cancer survival have been shown in England, as investigated by Lim et al.11 in 2006. They found that compared to winter, cancer diagnosis in summer and autumn was associated with improved survival. This is especially true in women with breast cancer and in both male and female patients with lung cancer. Importantly, in the same vein as our results, cumulative sunlight exposure in the months before cancer diagnosis was also a predictor of survival. In general, cancer survival is influenced more by seasonality in women than in men. For PCa, mortality was lower in the summer months when adjusting for sunlight exposure in the preceding 3 months. There are other reasons than sunlight exposure why the incidence of PCa may be season dependent. It has been shown that air pollutants such as ambient concentrations particulate matter 2.5 and nitrogen dioxide (NO₂) are associated with an increased risk of PCa.12 It is also possible that the temperature itself is a factor. In general, warmer climates and higher greenspace exposure have been associated with a decline in PCa incidence.13

The effect of reduced exposure to light on PCa could also be due to the effect of light on circadian rhythm. Studies suggest that circadian-related influences such as light may influence cancer susceptibility.14 Circadian genes influence different cancer-related biological pathways, including sex hormone regulation. Mancio et al.,15 in their systematic review on PCa and night shift work, found a weak (relative risk: 1.06, 95% CI: 1.01–1.12) increase in PCa risk in men with rotating night shifts only, but not in men on a fixed night schedule. This is in contrast to another meta-analysis that found that the risk for PCa in shift workers was 1.24 (95% CI: 1.05–1.46).8 This study found a dose–response relationship, meaning that the longer the shift work, the higher the PCa risk.

Seasonal changes in light exposure influence the endogenous clock in the suprachiasmatic nucleus (SCN). At the molecular level, desynchronization of circadian control of DNA replication, transcription, and cell metabolism may contribute to carcinogenesis. The repair of skin damage by sunlight is also dependent on sunlight, and the damage is greater in the morning when DNA excision repair is the lowest. One of the core clock genes is the core period (PER) clock factor, Per1. This factor is downregulated in PCa, inhibits transactivation of the androgen receptor (AR), and is induced by activated AR.13 Per1 inhibits growth in PCa cell lines and increases apoptosis. Activated AR stimulates Per1, which in turn attenuates AR activity to maintain hormonal homeostasis.16 In a large population-based case–control study of 1515 men, Wendeu-Foyet et al.17 reported an association between PCa and circadian gene variants such as retinoic acid-related orphan receptor alpha (RORA). RORA encodes a transcription factor that regulates the expression of several cancer-related genes and is associated with aggressive cancers.18 Sun exposure not only influences the incidence of cancers but also the treatment response.19,20,21 Outcomes in PCa have been shown to be worse when radiotherapy is administered later in the day, when sun exposure is lower. Hsu et al.22 found that radiotherapy administered after 5 p.m. was less efficient, especially in more aggressive cancers and in patients older than 70 years of age. These patients also had a higher incidence of gastrointestinal complications. The population in our study was very heterogeneous, and some patients were biopsied in private practice and some in a university hospital setting. Few biopsies are performed by residents in university hospitals. Therefore, we believe that the difference in Gleason scores is not due to sampling bias.

Our study has some limitations. Although a large number of patients were included, there is a risk of bias in our results. Our results suggest that the Gleason score could be influenced by other factors such as the presence or absence of general practitioners who refer patients for biopsy or urologists who might be absent during the summer months. However, we bundled 3 months together, which should have decreased the bias. Most of our patients were treated with brachytherapy, which might have skewed our data toward less frequent primary Gleason scores of 4.

In conclusion, we found that patients treated with radiotherapy were less likely to have primary Gleason score of 4 or 5 on biopsy when done in the 3 months following the highest sunshine exposure in our province. Correlation is not a proof for causation. Sunshine exposure can be used as a surrogate for many environmental causes associated with cancer.

AUTHOR CONTRIBUTIONS

DT analyzed the data and drafted the manuscript. GD drafted the manuscript. Both authors read and approved the final manuscript.

COMPETING INTERESTS

Both authors declare no competing interests.