Sex‐specific analysis of renal cell carcinoma histology and survival in Japan: A population‐based study 2004 to 2016

Abstract Background and aims In Western countries, sex differences in renal cell carcinoma (RCC) histology and survival have been previously described: compared with men, clear cell subtype is more common but overall prognosis is better among women. The goal of the present study was to examine sex differences in RCC histology and survival in Japan, using a large‐scale population‐based data set. Methods With the use of a population‐based cancer registry in Japan (2004‐2016), patients with primary RCC were followed for 5 years (median follow‐up time 2.1 years). We distinguished histological subtypes of clear cell, papillary, and chromophobe from “others” subtype. Sex‐specific prevalence ratio (PR) for each histological subtype was estimated by Poisson regression with robust variance, adjusted for age and year of diagnosis. Sex‐specific survival rates were estimated by Cox proportional hazard regression, adjusted for age, year of diagnosis, histological subtypes, and other prognostic variables, with multiple imputation. Results The prevalence of clear cell and “others” subtypes was similar between men and women among all the 5265 study subjects during the 12 years of study (clear cell, male 88.6% vs female 87.1%; “others“, male 5.3% vs female 5.3%). However, papillary subtype was less common among women than men (male 4.6% vs female 2.8%; PR = 0.63; 95% CI, 0.45‐0.88), while chromophobe subtype was more common among women (male 1.6% vs female 4.8%; PR = 3.18; 95% CI, 2.26‐4.47). Although “others” subtype (but not papillary/chromophobe subtypes) independently predicted prognosis (HR = 1.74; 95% CI, 1.32‐2.30), no sex differences were observed in RCC survival. Conclusion We did not observe a statistically significant difference in the prevalence of clear cell subtype between men and women in Japan, which differs from the pattern previously described in Western countries. Sex differences in RCC histology may not affect RCC survival in this population.


| INTRODUCTION
Renal cell carcinoma (RCC) ranks as the sixth most common type of cancer in men and 10th among women, worldwide, accounting for 5% and 3% of overall cancer incidence, respectively. 1 The incidence rate for RCC in Asian countries, including Japan, is roughly one-third compared with that in Western countries, 1 and the lower RCC incidence in Japan may be attributable to the lower prevalence of obesity (a prominent risk for RCC 2 ) in that country compared with the Western settings. However, the incidence has been rising in recent years in Japan. [1][2][3] Advanced imaging modalities, together with changes in lifestyle and behavioral risks (such as smoking, obesity, hypertension, and occupational stress), may underlie the increase in incidence. [3][4][5] Prognostic differences according to histological subtypes are well described. In the United States (USA), studies using the Surveillance, Epidemiology, and End Results (SEER) database suggest that the most common histological subtype is the clear cell subtype (approximately 80%-90%), followed by the papillary (approximately 4%-13%) and chromophobe (approximately 2%-5%) subtypes. [6][7][8] Compared with the clear cell subtype, patients with papillary and chromophobe subtypes tend to have better prognosis (hazard ratio [HR] approximately 0.67-0.98), but patients with other rare histological subtypes (eg, collecting duct and sarcomatoid differentiation) tend to have a poorer prognosis (HR approximately 1.81-2.21). 8 Previous single-center studies in the United States are consistent with this pattern. 9,10 More recently, studies in Western settings have reported sex differences in RCC histology and survival. [11][12][13][14] The clear cell subtype is likely more prevalent among women compared with men (88% vs 85%). 11 In addition, among all incident RCC, the papillary subtype tends to be approximately 0.5 to 0.6 times less prevalent, but the chromophobe subtype tends to be approximately 1.6 to 2.3 times more prevalent among women compared with men. 12,13 Sex differences in RCC histology, particularly in the clear cell subtype, may be partly attributed to lifestyle and behavioral factors such as obesity, 15 which is generally more prevalent in women than men across most countries, 16 and smoking. 17 However, compared with men, RCC prognosis has been reported to be better among women (HR 0.92) even though the clear cell subtype, which has a more unfavorable prognosis, was more prevalent among women. 11 Other female clinicopathological features (lower pathological grade, earlier stage of detection, and higher prevalence of chromophobe subtype) linked to favorable prognosis may underlie these paradoxical sex differences in RCC histology and survival. 8,[11][12][13] Reports of sex differences in RCC histology and survival remain scarce in Asian settings. For example, in Japan, although the Japanese Urological Association has reported descriptive statistics of RCC histology and 5-year overall survival with multicenter data (but not population-based data), 18 sex differences in RCC histology and survival have not been previously assessed. One study in South Korea, with 1508 RCC patients, examined sex differences and reported lower prevalence of clear cell subtypes among women compared with men (female 72% vs male 84%) but a nonsignificant sex difference in survival, which contradicts the Western pattern. 19 However, this was a relatively small, single-center study, and it was not population based.
Accordingly, the goal of the present study was to examine sex differences in RCC histology and survival in Japan, using a large-scale population-based data set with over 5000 RCC patients. We sought to examine whether sex differences exist in RCC histology and whether there are overall survival differences by sex and by histological subtypes in Japan. Also, we sought to determine whether a sex difference in survival, if there is one, persists even after controlling for histological subtypes and other potential prognostic factors such as pathological grade, tumor stage, treatment, socioeconomic status (SES), and smoking habits.

| Data setting and study subjects
A large, population-based data set (2004-2016) of Kanagawa Cancer Registry (KCR), a survey of over nine million people in Kanagawa prefecture that covers approximately 7% of the Japanese population, was used for analysis. Details of the study database have been previously described elsewhere. 20,21 Briefly, Kanagawa Prefecture, a metropolitan prefecture located next to Tokyo, is the second largest prefecture in Japan, and KCR is one of the largest local cancer registries in Japan.
Well-trained tumor registrars certified by the training program of the Japanese Association of Cancer Registries, whose program is accredited by SEER, are responsible for data collection. The data included basic information (sex, age, date of diagnosis, and date of death/last follow-up) and clinical information (pathology, stage, and treatment). Uniquely, KCR collected information on occupation and smoking behaviors, if available, during the study period (approximately 10% of the registered cases); however, these data are no longer  previous SEER studies. 6,8,22 The 5-year overall survival was defined by the right-censored, observation duration (person-years) from the date of initial diagnosis to the date of death/last follow-up (median follow-up time, 2.1 years).

| Covariates
The age (1-year age category) and year of diagnosis (calendar year) were adjusted as continuous variables across all statistical models. To  and smoking habits (never/ever).

| Statistical analysis
The background characteristics between men and women were compared by t test or chi-squared test.  Table 2). The PR of the "others" subtype did not differ between men and women ( Table 2).
In survival analyses, although the 5-year overall survival rate was 72% in this population, patients with the "others" subtype had a poor prognosis (30.3%, Table 1). However, no statistically significant differences were observed between men and women, even after stratifying by histological subtypes (Table 1). In the Cox regression analysis, although the chromophobe subtype predicted a better prognosis and the "others" subtype predicted a poor prognosis, only the "others" subtype predicted prognosis in the maximally adjusted model (HR 1.74; 95% CI 1.32-2.30, model 4, Table 3). However, we did not observe a statistically significant difference in the 5-year overall survival between men and women in model 1 through model 4 (Table 3).
In sensitivity analyses, the observed patterns were almost identical to the main results (  18 is likely at the lower end of published estimates compared with data from previous studies, [6][7][8]12,13,19 suggesting regional disparity in RCC histology.
Regarding sex disparities in RCC survival, women have been previously reported to have better prognosis compared with men, 11,14,19 which we did not observe in the present study. In the Western settings, better prognostic factors, including smaller tumor size, low pathological grade, and early stage, are likely more prevalent among women compared with men. 11,14,27 However, in this non-Western setting, we did not observe a sex difference in grade/stage. Similarly, a sex difference for grade/stage/tumor size was not found in South Korea. 19 The similar distributions of better prognostic factors between men and women might partly underlie the observed absence of sex disparity in RCC survival in our study, as opposed to studies in the Western setting. 11 Several limitations in this study should be noted. First, although our data set was population-based, it only represents approximately 7% of the Japanese population in one geographic region, and our obtained pathology diagnoses were not based on a central pathology review. In addition, other relevant outcomes (eg, relative survival) 28 were not evaluated, and complete data were limited for histology and other prognostic variables (including stage) because of missing data, thereby limiting internal and external generalizability. However, our sensitivity analyses with multiple imputation and complete data yielded almost identical results. Second, although we assessed SES and smoking habits, we could not assess other potential predicting factors such as metabolic disorders (eg, obesity, hypertension, and diabetes), tumor size, or performance status, as seen in previous studies. 1,6,10,11,19 However, hypertension and diabetes did not affect prognosis in a previous study in South Korea. 19 Third, because of the limitation of the data, we were not able to classify histological subtypes further (such as type 1 and type 2 papillary RCC). 28 Since patients with type 2 papillary RCC tend to have poorer prognosis, 29 future studies focusing on sex differences in papillary RCC subtypes are needed.
Despite these limitations, the strengths of our study included the size, as this is one of the most extensive studies conducted for evaluating sex differences in RCC histology in the non-Western setting.
Our distribution of the clear cell subtype (88.2%) estimated with population-based data was similar to the SEER data, 6,7,11,21 suggesting our reduced bias compared with nonpopulation-based studies. 18,19 In addition, while previous population-based studies did not include SES or smoking habits, 11 we were able to take account of these characteristics in our study.
Lastly, RCC patients with the "others" subtype tend to have a poorer prognosis compared with those with the clear cell subtype, 8,30 and we confirmed this disparity in Japan. In contrast to good prognoses in clear cell, papillary, and chromophobe subtypes (even though survivals tend to differ slightly among the three subtypes), 7 prognosis of the "others" subtype remains poor, with aggressive pathological features (eg, approximately 50% of RCC patients having the "others" subtype tend to present with metastasis). 8 Hence, further studies exploring effects of standard and novel agents for this high-risk population are warranted. 29 In conclusion, sex differences in RCC histology (papillary and chromophobe subtypes but not clear cell and other subtypes) appear to exist in Japan, which differs from the pattern previously described Japan Society for the Promotion of Science (JSPS KAKENHI JP18K17351). The supporting source had no involvement in study design; collection, analysis, and interpretation of data; writing of the report; and the decision to submit the report for publication.

CONFLICT OF INTEREST
The authors declare no potential conflicts of interest. have been explained.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the Kanagawa Cancer Registry. Restrictions apply to the availability of these data, which were used under license for this study by the Kanagawa Cancer Registry; research data used in the study cannot be made publicly available directly by the authors. If any person wishes to verify our data analysis, they are most welcome to contact the corresponding author.