Stage Evaluation of Cystic Duct Cancer
Article information
Abstract
Purpose
Cystic duct cancers (CDCs) have been classified as extrahepatic bile duct cancers or gallbladder cancers (GBCs); however, it is unclear whether their clinical behavior is similar to that of distal extrahepatic bile duct cancers (DBDCs) or GBCs.
Materials and Methods
T category of the CDCs was classified using current T category scheme of the GBCs and DBDCs, and clinicopathological factors were compared among 38 CDCs, 345 GBCs, and 349 DBDCs. We modified Nakata’s classifications (type 1, confined within cystic duct [CD]; combined types 2-4, extension beyond CD) and compared them.
Results
No significant overall survival (OS) difference was observed between the patients with CDC, GBC, and DBDC. The T category of GBC staging was more accurate at distinguishing OS in patients with CDC than the DBDC staging. Patients with T3 CDC and GBC showed a significant OS difference when using the T category for GBC staging, while those with T1-T2 CDC and GBC showed no significant difference. In contrast, the T category of DBDC staging did not show any significant OS difference between patients with T1-T2 CDC and DBDC or T3 CDC and DBDC. Patients with type 1 CDC had significantly better OS than those with combined types.
Conclusion
Unlike GBCs and DBDCs, CDCs exhibit distinct clinicopathological characteristics. The OS is better when the CDC confines within the CD, compared to when it extends beyond it. Therefore, we propose a new T category scheme (T1, confined to CD; T2, invaded beyond CD) for better classifying CDCs.
Introduction
The cystic duct (CD) connects the extrahepatic duct to the neck of the gallbladder (GB) and has a mean length of 3 cm and a diameter of approximately 2-3 mm [1,2]. Its epithelial lining cells are tall, columnar cells identical to those of the GB. Smooth muscle fibers of the GB are typically observed beneath the basement membrane of the epithelial cells and are loosely arranged in circular, longitudinal, and oblique manners. In contrast, fibrous tissue is observed underneath the basement membrane of epithelial lining cells of the CD [3].
Primary cystic duct cancer (CDC) is an extremely rare malignancy of the biliary tree, accounting for only 3% of bile duct cancers [4,5]. Farrar initially described the CDC in 1951 as a cancer restricted to the CD without associated malignancies in the GB, common hepatic duct (CHD), or common bile duct (CBD) with histopathologic confirmation [6]. However, Farrar’s strict criteria resulted in classifying “non-restricted CDC” as “non-CDC” when a cancer epicentered in the CD but spread to the surrounding CD structures [6]. CD was short and lack of a well-formed muscle layer, so cancer cells of CD can readily invade nearby structures, including the GB, CHD, and CBD [7,8]. Based on these findings, the prevalence of primary CDC could be higher when considering criteria other than Farrar’s original criterion. In 2003, Ozden et al. [9] proposed to classify CDC as a GB cancer (GBC). Nakata et al. [10] modified the definition of CDC in 2009 to include malignancies centered in the CD that invaded to the bile duct and/or GB and categorized CDCs into four types based on the extent of tumor spread.
CDCs had been alternatively staged as either extrahepatic bile duct cancers or GBCs in the previous editions of the American Joint Committee on Cancer (AJCC) cancer staging manual. The 6th edition of the AJCC cancer staging manual classified CDC as a subgroup of extrahepatic bile ducts cancers [11]. However, in its 7th edition, the classification of primary CDC as a subset of GBCs was established [12] and subsequently upheld in the 8th edition [13]. The clinicopathologic features of CDCs might be closer to those of distal extrahepatic bile duct cancers (DBDCs) than GBCs because the histological features of the CD resemble more closely to those of the extrahepatic bile duct than the GB [3]. However, our current understanding of CDC is limited due to its rarity and scarce literature. The objective of this study was to elucidate the clinicopathologic features and prognosis of CDC and further propose a more suitable staging system for primary CDC.
Materials and Methods
1. Case selection
Initially, 46 cases of surgically resected CDCs operated between 2012 and 2019 were selected from the records of the Department of Pathology at Asan Medical Center.
Epicenters of the tumors were evaluated by both gross and microscopic examinations. Gross images and preoperative radiologic images, including magnetic resonance imaging (MRI) and computed tomography (CT), of the surgically resected CDC cases were evaluated. After reviewing hematoxylin and eosin–stained slides, two patients with high-grade biliary intraepithelial neoplasias were excluded. Five and one patient whose tumor epicenters were the DBDC and pancreas, respectively, were excluded. Finally, 38 patients with surgically resected CDC were selected. Among the selected 38 patients, 24 had an R0 resection, while 14 had an R1 resection. In addition, we selected 345 patients with surgically resected GBCs between 2005 and 2012 and 349 with surgically resected DBDCs between 2008 and 2015 for clinicopathologic comparison with the CDC. The T and N categories of the CDCs, GBCs, and DBDCs were assessed according to the 8th edition schemes of the AJCC cancer staging manual, and their histologic types differentiation were classified according to the 5th edition of the World Health Organization (WHO) classification of digestive system tumors [13,14].
Data, such as age at diagnosis, sex, and survival outcomes, were retrospectively obtained from electronic medical records. Two pathologists conducted pathological examinations of the tumor extent and size, histologic type, differentiation, growth pattern, intracholecystic papillary neoplasm, lymphovascular invasion, perineural invasion, and depth of invasion.
2. CDC definition and classification
CDC was defined as cancer epicentered in the CD, as determined by gross and microscopic examinations. Cancers with epicenters in the GB, CBD, and CHD and extending into CD were excluded. CDC in this study was classified into four types based on the extent of spread proposed by Nakata et al. [10]: type 1 CDC, the CD contained the entire tumor; type 2 CDC, where the tumor reached the GB; type 3 CDC, where the tumor reached the CHD or the CBD, including extension into the lumen and external invasion to the bile duct wall; and type 4 CDC, where the tumor reached both the GB and bile duct simultaneously (Fig. 1).
Schematic drawing, and representative gross, radiologic, and microscopic images of cystic duct cancers based on original Nakata’s classification. CBD, common bile duct; CD, cystic duct; GB, gallbladder.
3. Statistical analysis
To analyze the distributions of the variables among the CDCs, GBCs, and DBDCs, Pearson’s χ2 test or Fisher’s exact were applied for categorical data. The overall survival (OS) was defined as the time from surgical resection to death or the last follow-up examination. The OS rates were calculated using the Kaplan-Meier method. Associations between survival rates and various clinicopathologic factors were assessed using the log-rank test. CDC cases were divided into T1-2 and T3-4 groups using the 8th edition schemes of the T category of the AJCC cancer staging manual for GBC or DBDC, respectively. The log-rank test was then used to compare the survival times of each group with the same T categories between the CDCs and GBCs and between the CDCs and DBDCs. All statistical analyses were performed using R software (https://www.R-project.org), and all p-values < 0.05 were considered statistically significant.
Results
1. Patient characteristics
The clinicopathologic characteristics of the 38 CDCs are summarized in Table 1. The detailed clinicopathologic features of CDC patients are provided in S1 Table. The mean patient age was 67.4±9.4 years, male to female ratio was 1.2, and mean tumor size was 28.6±14.1 mm. Based on Nakata’s classification, there were five type 1 (13.2%), two type 2 (5.3%), 20 type 3 (52.6%), and 11 type 4 CDCs (28.9%), respectively. There were 36 tubular adenocarcinomas (94.8%), one mucinous carcinoma (2.6%), and one mixed adenocarcinoma-neuroendocrine carcinoma (2.6%). Fifteen (40.5%) were well-differentiated, 19 (51.4%) were moderately differentiated, and three (8.1%) were poorly differentiated cancers. Lymphovascular and perineural invasions were noted in 17 (44.7%) and 27 (71.1%) cases, respectively. Lymph node metastasis was observed in 37.8% (n=14) of cases. Seventeen patients (44.7%) had simple or extended cholecystectomy with or without bile duct resection, 10 (26.3%) had pylorus-preserving pancreaticoduodenectomy, nine (23.7%) had cholecystectomy with bile duct resection and partial hepatectomy/segmentectomy, and two (5.3%) had Whipple’s operation. Among the 38 CDCs, 10 were CDCs associated with intraductal papillary neoplasms (IPNs), of which eight (80.0%) cases were well-differentiated cancers, and two (20.0%) had perineural invasion. Patients with IPNs (median OS, 30.5 months) tended to have a better OS than those with conventional CDCs (26.0 months, p=0.172) (S2A Fig.). On the other hand, subgroup analysis with R1 status showed no significant OS difference between conventional CDC (median, 17.0 months) and CDC with IPNs (30.6 months, p=0.547) (S2B Fig.). Of the 24 R0 resected CDCs, five cases (20.8%) were associated with IPNs. In contrast, of the 14 R1 resected cases, five cases (35.7%) associated with IPNs. Patients with IPNs and R0 resection had significantly better OS (median, 49.0 months) than those with conventional CDCs and R0 (28.0 months, p=0.049) (S2C Fig.). None of the patients showed distant metastasis at the time of diagnosis. The median follow-up period was 28.0 months (range, 6 to 98 months).
Clinicopathologic characteristics of 38 CDCs
2. Application of GBC and DBDC T category scheme
All CDC cases were restaged using the T category scheme of GBC and DBDC staging (Table 2). Per the GBC staging system, 31 CDC cases with CBD invasion (81.6%) were classified as T3. CHD invasion was observed in 20 cases, liver invasion was observed in three cases, and pancreatic invasion was observed in two cases.
Comparison of the application of GBC and DBDC T category schemes to CDC cases
Per the DBDC staging system, there were 13 T1 (34.2%), 23 T2 (60.5%), and two T3 (5.3%) cancers. Comparing the applications of GBC and DBDC T category schemes, nine T3 cases were downgraded to T1 classification, and 20 T3 cases were downgraded to T2.
3. Preoperative radiologic findings
Preoperative CT and/or MRI data were available for 37 CDC cases. In most CDC cases, they were demonstrated as circumferential enhancing wall thickening or enhancing mass on CT and MRI (Fig. 1).
4. Comparisons of clinicopathologic factors according to the Nakata’s classification
Comparisons of clinicopathologic factors of the CDCs according to the Nakata’s classification are shown in Table 3. When the T category scheme of the GBC staging was employed, the T1 and T2 categories were common in types 1 cancers and the T3 category was more common in types 3 and 4 CDCs (p < 0.001). Perineural (p=0.045) and lymphovascular (p=0.034) invasions were more common in other types (types 2-4) of CDCs than in type 1 CDCs. We modified the Nakata’s classification to compare type 1 and combined types 2-4. The T category scheme of GBC staging showed that all type 1 CDCs were categorized as T1-T2 cancers, while most combined types 2-4 CDCs were classified as T3 cancers (p < 0.001). Perineural (p=0.019) and lymphovascular (p=0.053) invasions were more common in combined types 2-4 CDCs than in type 1. CDCs associated with IPN were more commonly observed in type 1 CDCs than in the combined types 2-4 CDCs (p=0.012).
Correlations between original and modified Nakata’s classification and various clinicopathological characteristics of CDC
5. Comparison of clinicopathologic characteristics between CDCs and GBCs and between CDCs and DBDCs
The clinical characteristics of CDC, GBC, and DBDC are compared in Table 4. The mean age of the patients with CDC (mean, 67.4 years) was significantly older than those with GBC (62.9 years, p=0.011). CDCs had more frequent lymphovascular (p=0.020) and perineural (p < 0.001) invasions than GBCs. The CDC cases had more differentiated cancers than the DBDC cases (p=0.006).
Comparisons of clinicopathologic characteristics among CDC, GBC, and DBDC
T3 tumors were more frequent in CDCs than in GBCs (p < 0.001) per the T category scheme of the GBC staging. Using the T category scheme of the DBDC staging, the CDCs had more T1-T2 cancers than DBDCs (p=0.027).
6. Survival analysis of CDC using T category schemes of GBC and DBDC staging systems
Survival analysis was conducted on patients with CDC, utilizing the GBC and DBDC T category staging systems, to ascertain the most suitable T category for clinical use. The T1 and T2 cases were grouped and their survival rate was compared to those of T3 cases (Fig. 2). When the OS was compared according to the T category of the GBC scheme, a significant difference in the OS was observed (p=0.029) (Fig. 2A). When the OS was compared according to the DBDC scheme, the median OS time in the patients with T1-T2 and T3 CDCs were 26.0 and 43.0 months, respectively. No significant survival difference was observed when the T category of the DBDC scheme was applied (p=0.779) (Fig. 2B).
The overall survival (OS) comparisons according to cystic duct cancer (CDC) according to the T category of gallbladder cancer (GBC) (A) and distal extrahepatic bile duct cancer (DBDC) (B) schemes. (A) When the OS was compared according to the T category of the GBC scheme, the median OS time in CDC patients with T3 was 24.0 months. In contrast, those with T1-T2 CDC did not reach the median OS (5-year OS rate, 57.1%). A significant OS difference was observed (p=0.029). (B) When the OS was compared according to the T category of the DBDC scheme, the median OS times in CDC patients with T1-T2 and T3 categories were 26.0 and 43.0 months, respectively. No significant survival difference was observed (p=0.779).
7. Comparison of survival analysis of CDC between GBC or DBDC
We compared the OS among patients with CDC, GBC, and DBDC based on the T1-T2 and T3 categories of the GBC and DBDC (Fig. 3). When the GBC scheme was applied, the OS rates were not different between patients with T1-T2 GBC and T1-T2 CDC (p=0.721) (Fig. 3A). However, patients with T3 CDC had a significantly better OS than those with T3 GBC (p=0.008) (Fig. 3B). When the DBDC scheme was implemented, the OS rates were not different between patients with T1-T2 DBDC and T1-T2 CDC (p=0.168) (Fig. 3C). In addition, no significant OS difference was observed between patients with T3 CDC and T3 DBDC (p=0.522) (Fig. 3D).
The overall survival comparisons among cystic duct cancer (CDC), gallbladder cancer (GBC), and distal extrahepatic bile duct cancer (DBDC) patients based on the T1-T2 and T3 categories of the GBC and DBDC scheme were applied. (A) When the GBC scheme was applied, overall survival rates were not different between patients with T1–T2 GBC and T1-T2 CDC (p=0.721). (B) However, patients with T3 CDC had significantly better overall survival than those with T3 GBC (p=0.008). When the DBDC scheme was implemented, the overall survival rates were not different between patients with T1-T2 DBDC and T1-T2 CDC (p=0.168) (C) and between patients with T3 CDC and T3 DBDC (p=0.522) (D).
All 30 CDC patients were categorized as T3 per the GBC staging scheme but were categorized as T1-T2 per the DBDC staging scheme. We grouped these 30 patients as the debated CDC group and compared OS with patients with GBC and DBDC and their T category subgroups (T1-T2 and T3) (Fig. 4). The 5-year OS of the 30 patients with debated CDC group (28.4%) was worse than those with T1-T2 DBDC (48.8%, p=0.06) but better than those with T3 GBC (14.6%, p=0.006). In addition, patients with T1-T2 GBC had a better OS than those with T1-T2 DBDC (5-year OS, 64.5% vs. 48.8%, p=0.003).
The overall survival comparisons of 30 patients with debated cystic duct cancer (CDC), T1-T2 and T3 gallbladder cancers (GBCs), and T1-T2 and T3 distal extrahepatic bile duct cancers (DBDCs). The overall survival of the patients with 30 debated CDC group (5-year overall survival, 28.4%) was worse than those with T1-T2 DBDC (48.8%, p=0.065) but better than those with T3 GBC (14.6%, p=0.006). In addition, patients with T1-T2 GBC (64.5%) had a better overall survival than those with T1-T2 DBDC (48.8%, p=0.003).
8. Survival analysis of CDC classified per Nakata’s classification
The patients with type 1 CDC did not reach the median OS (5-year OS rate, 60%). In contrast, the median OS times in those with type 2, 3, and 4 CDCs were 17, 29.5, and 14 months, respectively. Patients with type 1 CDC tended to have better OS than those with other types of CDC (p=0.056, overall comparison) (Fig. 5A). A tendency of OS difference was observed between types 1 and 2 CDCs (p=0.228), types 1 and 3 CDCs (p=0.060), and types 1 and 4 CDCs (p=0.060) in a pair-wise comparison. However, no significant OS difference was observed between patients with types 2 and 3 (p=0.399), types 2 and 4 (p=0.938), and types 3 and 4 (p=0.690) CDCs. Therefore, we grouped types 2, 3, and 4 into “combined types” (combined type 2-4) and compared their OS with type 1 CDC. The median OS time in combined type 2-4 CDC was 22 months. Patients with type 1 CDC had significantly better OS than those with combined types (p=0.010) (Fig. 5B).
The overall survival comparisons of patients with cystic duct cancer (CDC) by the original (A) and modified (B) Nakata’s classification. (A) Patients with type 1 CDC had a tendency toward better overall survival than those with other types (p=0.056, overall comparison). A tendency of overall survival difference was observed between types 1 and 2 CDCs (p=0.228), types 1 and 3 CDCs (p=0.060), and types 1 and 4 CDCs (p=0.060) by pair-wise comparisons. However, no significant overall survival difference was observed between patients with types 2 and 3 (p=0.399), types 2 and 4 (p=0.938), and types 3 and 4 (p=0.690) CDCs. (B) Cases with types 2, 3, and 4 CDCs were merged into a combined group, and its overall survival compared with type 1 CDC. Patients with type 1 CDC had significantly better overall survival than those in the combined group (p=0.010).
Discussion
The key findings of CDCs in this study are as follows: The mean cancer size was approximately 3 cm. Tubular adenocarcinomas, either well- or moderately differentiated, were most common. Lymph node metastasis was observed in one-third of the cases. CDCs had more T3 tumors with the GBC scheme and more T1-T2 tumors with the DBDC scheme. Applying the T category scheme of GBC instead of the DBDC classification scheme improved the OS discrimination in patients with CDC, although patients with T3 CDC had a better OS than those with T3 GBC. Patients with type 1 CDCs had better survival than the combined types 2-4.
Only a few previous studies have compared CDC, GBC, and DBDC to evaluate their clinicopathological features and prognosis [15,16]. Nakanishi et al. [15] compared 47 CDCs with 199 perihilar extrahepatic bile duct carcinomas (PBDCs) and 44 GBCs, and 47 CDCs (100%) were classified as pT3 or pT4 per the 8th edition of the AJCC GBC T classification scheme. The OS of the patients with CDC was similar to that of the 27 patients with pT3 GBC. When the patients with CDC were classified per the 8th edition of the AJCC of PBDC scheme, the majority of CDCs (36 cases, 77%) were classified as pT2 cancers; however, the OS in patients with CDC was significantly worse than that of those with pT2 PBDC (29 vs. 48 months, p=0.040) [15]. Nakanishi et al. [15] suggested that the clinicopathologic behavior of CDCs differs from that of PBDCs.
Yu et al. [16] evaluated CDCs and compared them with GBCs and extrahepatic cholangiocarcinomas using the Surveillance, Epidemiology, and End Results (SEER) 18 registries database. The mean tumor size was largest for GBCs compared to CDCs and DBDCs (mean, GBC, 3.4 cm; DBDC, 2.7 cm; and CDC, 2.8 cm) in the present study, which is consistent with the observation of Yu et al. [16] (mean, GBC, 3.3 cm; DBDC, 2.5 cm; and CDC, 2.5 cm), likely due to anatomical features. Due to its pouch-like shape, the GB allows for more tumor growth than cystic or extrahepatic ducts, which have less space for tumor expansion [3]. In the present study, the lymphovascular and perineural invasion was observed in 45.7% and 71.1% of CDCs, respectively, which were similar to those of the DBDCs (42.4% and 79.1%) and more than those for GBCs (25.3% and 27.1%) [17,18]. One recent GBC study compared the clinicopathologic behavior of GBCs based on their location (fundal end versus CD end), and patients with GBCs in the CD end had a worse recurrence-free survival than those with GBCs in the fundal end [19]. The findings of the current and previous studies may be accounted for by the presence of numerous lymphatic, vascular, and nervous structures in the CD, which is close to the hepatic hilum [20,21].
We observed that patients with T3 CDC had a better OS than those with T3 GBC. This could be explained by the fact that serosal invasion or adjacent organ invasion is currently classified as the T3 category in GBCs. Nevertheless, CDCs located in proximity to the GB and CBD facilitate the lateral spread of cancer cells. Therefore, even if the CDC is classified as T3, it may be less advanced than GBC. In addition, in the current T category scheme of DBDC, adjacent organ invasion is not considered. Consequently, in our study, the majority of CDC patients were categorized as T1 or T2 when the DBDC staging was applied, and those patients with CDC exhibited a generally unfavorable OS compared to the patients with T1-T2 DBDC. One explanation for these findings is that the CD, which has a diameter of 2-3 mm with a thin layer of fibrous tissue beneath the basement membrane, predominantly enables lateral spread to nearby structures before cancer cell invasion into the deeper level. This anatomical characteristic may limit the invasion depth of cancer cell invasion to 12 mm, which separates the T2 and T3 categories in the DBDC staging system. As a result, CDCs that would have been classed as T3 under the GBC scheme due to direct extension to neighboring organs may not surpass the 12 mm limit invasion depth and would be classified as T2 in the DBDC scheme. This could explain the observed differences between the GBC and DBDC staging schemes. We categorized this group as T3 with GBC staging and T1-T2 with the DBDC staging and compared them to the T1-T2 and T3 categories of patients with GBC and DBDC and this group had an OS between T1-T2 DBDC and T3 GBCs (Fig. 4). In addition, we observed that CDCs had more T3 tumors than GBCs and more T1-T2 cancers than DBDCs. The T category scheme of the GBC distinguished OS in patients with the CDC more accurately than the DBDC classification scheme. However, patients with T3 CDC had a better OS than those with T3 GBC. These results support the necessity for an innovative staging system for CDCs.
There are several proposals for CDC classifications [10,22,23]. We compared CDCs with Nakata’s classification and observed that type 1 CDCs (CDC confined within CD) had less lymphovascular and perineural invasion than other types (types 2-4; CDC invasion beyond CD). Additionally, patients with type 1 CDC had significantly better survival rate than those with other types (types 2-4). The R1 resection rate for type 1 was 20% (1/5), compared to 39% (13/33) for the other types. The higher incidence of R0 resections in type 1 may be attributed to the tumor’s size, location, and morphological characteristics, which could account for the more favorable prognosis observed in patients with type 1 CDC. Notably, CDCs associated with IPN comprised 40% (4/10) of type 1 CDCs, which corresponded to 80% (4/5) of all type 1 CDCs, and the patients with CDC and IPN had a tendency to have a better OS than those with conventional CDC. These results suggest that modified Nakata’s classification (T1, confined to CD versus T2, invasion beyond CD) can be used as a novel T category for CDC (Fig. 6).
Proposed schematic representation of the new system for cystic duct cancer. (A) pT1: confinement in cystic duct. (B) pT2: invades adjacent organ or structure beyond the cystic duct.
The strength of the present study is that, through various statistical analyses, our study found that neither the current GBC nor the DBDC staging system appropriately staged CDC. We aimed to propose a new criterion that better reflects patient prognosis. Furthermore, since CDC frequently involves CBD and CHD, extended cholecystectomy and bile duct resection may improve patient survival rates, considering clinical and radiological information. Our study has several limitations. The majority of CDC cases fell into the T3 category per GBC staging and the T1-T2 category per DBDC staging schemes; making subgroup analyses comparing other T categories was challenging. In addition, the small number of CDC cases comparing GBC or DBDC limits statistical power.
In summary, this study showed that CDCs have distinct clinicopathological characteristics compared to GBCs and DBDCs. Therefore, creating a unique T category scheme tailored to CDC rather than using the T category scheme of GBC may be preferable. We propose new T category scheme (T1, confined to CD; T2, invaded beyond CD) for better classifying CDCs.
Electronic Supplementary Material
Supplementary materials are available at Cancer Research and Treatment website (https://www.e-crt.org).
Notes
Ethical Statement
This study was approved by the Institutional Review Board (IRB) of Asan Medical Center with a waiver of informed consent (approval No. 2020-1205).
Author Contributions
Conceived and designed the analysis: Kim Y, Lee W, Hong SM.
Collected the data: Kim Y, Sung YN, Lee KJ, Yoo D, Jun SY, Lee W, Hong SM.
Contributed data or analysis tools: Jung H, Hwang S, Hong SM.
Performed the analysis: Kim Y, Jung H, Yoo D, Cho H, Hong SM.
Wrote the paper: Kim Y, Jun SY, Hwang S, Lee W, Hong SM.
Conflict of Interest
Conflict of interest relevant to this article was not reported.