Intraductal Papillary Neoplasm of the Bile Duct: Cytomorphologic and Molecular Features

INTRODUCTION

Intraductal papillary neoplasm of the bile duct (IPNB) is a rare, premalignant neoplasm of the bile duct that accounts for approximately 10% to 15% of all bile duct tumors although the prevalence may be higher in eastern Asia.^1,2 IPNB can progress from low-grade to high-grade dysplasia to eventual invasive adenocarcinoma. The lesion was first termed papillary carcinoma of the bile duct by Albores-Saavedra,³ and the term IPNB has been adopted and defined by the WHO Classification of Tumors since 2010.^4,5 IPNB grows within the bile duct and grossly appears papillary (villous) or nodular.⁵ Histologically, the lesion shows fibrovascular stalks or well circumscribed nodules lined by neoplastic epithelium. Based on the degree of cytologic atypia, IPNB can be further classified as IPNB with low-grade dysplasia and IPNB with high-grade dysplasia. When invasion is present, the term IPNB with associated invasive carcinoma is used. Although it has not been officially recognized as a distinct entity in the bile duct yet, emerging studies show that, as described in the pancreas, intraductal tubulopapillary neoplasm (ITPN) is another rare premalignant lesion that can also involve the bile ducts.^6–8 Histologically, ITPN is characterized by intraductal tubulopapillary growth.

It is important to recognize these uncommon lesions preoperatively since IPNB/ITPN, even with presence of associated invasive carcinoma, are prognostically better than conventional cholangiocarcinoma.^9,10 Noninvasive IPNBs can be cured by complete surgical resection, especially low-grade IPNBs. Cytology specimens are frequently obtained to evaluate lesions in the bile duct system using techniques such as brushing and fine-needle aspiration (FNA); therefore, pathologists may be able to diagnose these lesions at least in some cases but not in all cases, especially when the cellularity is limited. While cytology specimens are valuable, they can be diagnostically challenging due to the small sample size, loss of tissue architecture, and reactive changes that can produce atypia mimicking neoplasia. Integrating the clinical and radiologic impression as well as ancillary tests including molecular results becomes critical when interpreting these cytology specimens.¹¹ To date, there have only been a few case reports and an abstract of the cytomorphologic findings of IPNB.^12–15 The purpose of this study was to expand upon the current limited data to better elucidate the cytomorphology findings of this rare neoplasm to facilitate an accurate diagnosis. Herein, we presented 13 cases of IPNB, including 4 ITPNs, and described their clinicopathologic characteristics, imaging findings, cytomorphologic features, histologic correlation, and KRAS mutation results.

MATERIAL AND METHODS

RESULTS

DISCUSSION

IPNB is a precursor lesion in the biliary system which has the potential to progress to invasive adenocarcinoma.¹ Previous studies have described the clinical and histopathologic features of IPNB.^18–21 The epithelium of IPNB can show biliary, intestinal or gastric differentiation based on the morphologic and immunophenotypic features.^22–24 It is common to have more than one type of epithelium seen in an individual tumor. The tumor with predominant biliary epithelium appears to have worse prognosis as compared to that with either gastric or intestinal epithelium.^25–27 Interestingly, the IPNBs of biliary type are most frequently seen in North American and European countries while the intestinal and gastric type IPNBs are the predominant tumors in Ascian countries. Despite well characterization of IPNB on surgical specimens, the literature on the cytomorphologic features of these lesions is extremely limited. In fact, there are only a few case reports on IPNB cytomorphology.^12–14 This is the first study with a cohort of 13 cases that systemically evaluated the morphologic features in cytology specimens and correlated cytologic features with histopathologic follow-up. We believe that the findings in our study would help enhance the ability of cytopathologists to recognize this uncommon entity and therefore provide a more accurate diagnosis on limited cytology samples. The patients with IPNB in our study cohort were mostly middle-aged or elderly individuals with a male predominance, had mass or mass-like lesions with intraductal growth on imaging studies, and presented with obstructive jaundice or cholangitis-related symptoms, similar to those previously described.^5,28,29 Knowing these clinical and imaging characteristics may help increase the prospect of rendering an accurate diagnosis based on the cytomorphologic findings.

The most prominent cytomorphologic feature, identified in more than three-quarter cases in our cohort, was the cellular specimens with papillary or/and complex papillary clusters, which reflects the eponymous, exophytic papillary projections into the lumen of the bile ducts amenable to cytology sampling techniques.^30,31 This unique cytomorphologic feature of IPNB was more likely seen in FNA (7/7, 100%) than (4/7, 57%) brushing samples. Papillary architecture with fibrovascular cores was best seen on the cell block sections. In general, the background was clean with extracellular mucin, acute inflammation, and necrosis seen in only 4 (31%), 2 (15%), and 4 (31%) cases, respectively. There are no obvious differences in papillary architecture between low-grade IPNB vs high-grade IPNB/adenocarcinoma. However, some cytomorphologic features such as crowded groups, dyscohesive groups, and single cells are usually absent in low-grade IPNBs. Other features such as anisonucleosis and necrosis are also not seen in low-grade IPNBs. At the cellular level, the tumor was composed of mixed columnar, tall columnar and cuboidal cells; however, columnar cells were the predominant cell type in 11 out of 13 cases (85%). Another notable cytomorphologic feature is that tumor cells had relatively smooth nuclear membranes and fine granular chromatin; it would be interesting to compare the cytomorphology side by side with cholangiocarcinoma, the most common malignancy in the bile ducts. Interestingly, ITPN looks slightly different from IPNB on cytology. Architecturally, ITPN sometimes had clusters of epithelial cells with cribriform/tubular arrangement, which was not seen in IPNB. The tumor cells of ITPN were relatively uniform in term of nuclear to cytoplasmic ratio. In contrast, the tumor cells of IPNB showed cytological atypia ranging from low- to high-grade except for low-grade IPNB that showed only low-grade atypia. However, there was only 4 ITPN cases in our cohort; the case number is too small to generalize these findings.

The current six-tier diagnostic schema for pancreatobiliary cytology specimens is based on the guidelines of the Papanicolaou Society of Cytopathology for Reporting Pancreaticobiliary Cytology (PSC-RPC).³² Although IPNB may be distinctive when overtly papillary fragments are present, an accurate diagnosis is likely to be challenging on cytology specimens since our cohort demonstrated 3 out 13 (23%) cases with an atypical cytological diagnosis. Challenges include the lack of well-defined published criteria based on cytology. According to the PSC-RPC guidelines, the cytologic features of IPNB were extrapolated from the histopathological features.³³ A second inherent challenge in diagnosis is that some cytology samples are limited by cellularity and/or preservation artifacts. For example, 2 of 3 atypical diagnoses in our cohort had moderate cellularity and one case was originally called reactive epithelial cells due to low cellularity and preservation artifacts (Figure 5). Thirdly, brushing technique, the most used sampling technique for bile duct lesions, has known limitations.^30,31 Brushing may result in cellular specimens for IPNB lesions; however, we observed fewer specimens of this type contained the entity-defining large tissue fragments (papillary or complex papillary clusters). Three out of four cases with atypical or reactive diagnosis were brushing samples. Lastly, IPNB is an uncommon entity with diverse of architectural (papillary and tubulopapillary) and morphologic (columnar, tall columnar and cuboidal cells) features.

Surgical resection is advised for suspected IPNB regardless of grade because high-grade dysplasia and invasion are seen in a high proportion of resected neoplasms and furthermore, these obstructive lesions can lead to symptomatic cholangitis and jaundice.²¹ Nonetheless, accurate recognition of high-grade cytological atypia remains critically important for providing an accurate pre-operative assessment of malignancy risk. The cytologic features defining high-grade dysplasia or adenocarcinoma for pancreaticobiliary lesions are documented by other sources.^33–35 A unique feature seen in IPNB/ITPN cases was the columnar/tall columnar cells with enlarged nuclei with a papillary or complex papillary architecture. We described the cytological features seen in the few low-grade neoplasms in this cohort, but there were too few cases of resected low-grade IPNB to draw meaningful conclusions regarding cytomorphological differences between low-grade versus high-grade dysplasia versus invasive carcinoma.

It is well documented that tumors arising in the biliary tract are associated with recurrent molecular alterations.^21,36–38 Identification these molecular alterations may have potential diagnostic and therapeutic implications. Among them, KRAS mutation is probably one of the most recurrent molecular events. By demonstrating KRAS mutations in 44% of tested cases (4 of 9 cases) in this cohort, we found that the frequency of KRAS mutation in our study is comparable to a prior IPNB study.³⁷ In contrast, KRAS is mutated in up to approximately 90% of pancreatic ductal adenocarcinoma cases^39,40 and between 5 and 54% of conventional intrahepatic cholangiocarcinomas.^41–43 Interestingly, in our cohort, positive KRAS mutation was identified only in IPNB not ITPN cases. KRAS was wild type and unknown in 3 and 1 ITPN cases, respectively, consistent with the finding of a previous study.⁶ In addition to KRAS mutations, several other molecular alterations have been detected; however, no common mutation has been identified across all the subtypes. Schlitter et al. performed a multicenter study examining molecular alterations in 45 patients with IPNB and found that low-grade intraepithelial neoplasms showed mutated KRAS, overexpression of TP53 and loss of p16, while the later steps in tumor progression showed loss of SMAD4.³⁶ Yang et al. examined 37 IPNB cases and found that the most common mutation was KRAS (49%), followed by GNAS (32%), among others.³⁷ Aoki et al. examined 35 IPNB cases and found that the most common mutation was TP53 (34.3%), followed by STK11 (25.7%), KRAS (24%), among others.⁴⁴ In 3 of our cases that were submitted for comprehensive molecular testing, additional molecular alterations were found, including mutations in TP53, CTNNB1, SMAD4 and other genes. Thus, KRAS analysis may be helpful in detecting intraductal neoplasms at this site, including low-grade IPNB and positive KRAS mutation may favor IPNB over ITPN. Elucidating the molecular progression of IPNB may help diagnosis in cases with an indeterminate cytology and may potentially guide patient management options including target therapies.

We acknowledge that there are limitations to this retrospective study. We only included cases with concurrent/subsequent surgical pathology which incurs selection bias. The diagnoses of cytology specimens might be influenced by the results of concurrent biopsies. Therefore, we cannot provide data on the performance characteristics for the diagnosis of IPNB by cytology. Nonetheless, we provide detailed descriptive cytologic features which will foster a greater understanding of the cytology features of IPNB. In the presence of appropriate clinical and imaging findings, a sufficient cellular cytology sample may be accurately classified. In addition, our cohort was small, including only 9 IPNB and 4 ITPN cases. Although IPNB cases showed various epithelial subtypes on biopsy/resection specimens, it is difficult to correlate and generalize specific cytomorphologic features and molecular findings as to any specific epithelial subtype.

In conclusion, this study demonstrated that the cytomorphologic features of IPNB are heterogenous, yet a high proportion of specimens were cellular with papillary or compact clusters of columnar cells with vacuolated cytoplasm, enlarged nuclei, and granular chromatin. We emphasize that confirming a papillary neoplasm in the bile duct has clinical impact even if only the cytologic atypia present appears low-grade. Our review of the KRAS mutation status of these patients shows mutations were detected in both low- and high-grade neoplasms; however, at a lower frequency compared to pancreatic ductal adenocarcinoma.