Clinical Outcomes and Molecular Profiling of Pancreatic Acinar Cell Carcinoma: A Retrospective Study

Abstract

PURPOSE

Pancreatic acinar cell carcinoma (PACC) is a rare and aggressive form of pancreatic cancer that originates in the acinar cells of the exocrine pancreas. In this study, we aimed to investigate the clinical and molecular characteristics of patients with PACC at our institution.

METHODS

This was a retrospective study of patients with PACC seen at Mayo Clinic between 2002 and 2023. Baseline patient characteristics, tumor pathology, treatment strategies used, and survival outcomes were analyzed. Kaplan-Meier curves were estimated using newsurv macros in SAS.

RESULTS

The study included a total of 65 patients with PACC. The median age at diagnosis was 66 years. Almost half of the patients (48%) presented with resectable/borderline-resectable disease (n = 28). Five-year overall survival (OS) for resectable/borderline-resectable, locally advanced/unresectable, and metastatic disease were 72.0%, 21.6%, and 20.9%, respectively. Somatic and germline next-generation sequencing identified numerous potentially actionable targets including homologous recombination (43% somatic, 33% germline), RAF alterations (29% somatic), and mismatch repair (14% somatic).

CONCLUSION

Our findings underscore the heterogeneity and aggressive nature of PACC. Despite the improved prognosis for patients with resectable/borderline-resectable disease, OS remains poor, particularly for those with locally advanced or metastatic disease. The identification of actionable molecular targets in a significant proportion of patients highlights the potential for personalized therapeutic approaches. Future research should focus on tailored treatment strategies to exploit these molecular vulnerabilities, which may offer new options for improving outcomes in this rare malignancy.

INTRODUCTION

Pancreatic acinar cell carcinoma (PACC) is an uncommon neoplasm arising from the acinar cells of the exocrine pancreas. It represents a distinct entity from the more prevalent pancreatic ductal adenocarcinoma (PDAC) and encompasses approximately 1%-2% of all pancreatic malignancies.^1,2 The clinical manifestation of PACC often encompasses a spectrum of symptoms, ranging from nonspecific abdominal discomfort, weight loss, jaundice, and digestive complications. Certain factors such as smoking, occupational exposure to chemicals, and potential genetic predispositions have been postulated to contribute to its development.³

CONTEXT

• Key Objective

• How does pancreatic acinar cell carcinoma (PACC) differ from other pancreatic malignancies in its clinical outcomes, histopathological features, and molecular characteristics, and how can this inform treatment decisions?

• Knowledge Generated

• In patients with resectable PACC, outcomes were more favorable compared with pancreatic ductal adenocarcinoma, with a median recurrence-free survival of 36 months and a 5-year survival rate of 72%. Almost half (43%) of the patients tested for somatic and germline alterations exhibited actionable mutations.

• Relevance

• The optimal perioperative regimen for patients with PACC has yet to be determined. However, our findings underscore the importance of universal next-generation sequencing to inform treatment choices involving targeted therapies for select patients.

Arising primarily in adults around the age of 60 years, PACC carries a poor prognosis with a median overall survival (OS) of around 11-25 months.^4-7 Microscopically, these cells mirror normal acinar cells but exhibit uncontrolled proliferation and atypical cellular features.⁸ They typically form well-differentiated, cohesive glandular structures, often displaying prominent acinar differentiation with zymogen granules and an eosinophilic cytoplasm.^9,10 While they may maintain some degree of acinar cell differentiation, they also exhibit various architectural patterns, including solid growth, trabecular arrangements, and pseudopapillary structures, which can contribute to the tumor's heterogeneous appearance.^1,3 Immunohistochemical (IHC) staining often reveals the expression of acinar cell markers such as trypsin, chymotrypsin, and lipase, further confirming the acinar differentiation.¹⁰ Additionally, these tumors may exhibit varying degrees of cellular atypia and nuclear pleomorphism, features that can aid in their distinction from other pancreatic malignancies.¹¹

Management strategies for PACC involve a multidisciplinary approach and are mostly extrapolated from PDAC guidelines given the rarity of the condition.¹² In this study, we retrospectively reviewed patients with PACC at Mayo Clinic with the aim to better define the clinical and molecular characteristics of these tumors.

METHODS

We reviewed patients' records from Mayo Clinic (Arizona, Florida, and Minnesota) between 2002 and 2023 with diagnosis of PACC. Records were extracted from the pathology slides database as well as the Mayo Data Explorer. The reference pathologists (R.P. and C.H.) reviewed pathology results to confirm acinar differentiation. Those without clinical history or sufficient pathology information were excluded from survival analysis. Information on demographics, treatment courses, and next-generation sequencing (NGS) were collected. Analyses were conducted using SAS version 9.04. Continuous variables were summarized using mean, standard deviation, median, IQR, and range. Categorical variables were described using frequencies and proportions. Kaplan-Meier curves were estimated using newsurv macros in SAS. The study was approved by the Mayo Clinic institutional review board.

RESULTS

Patient Outcomes

A total of 65 patients with PACC were identified. Most of the patients were male (74%, n = 48), and the median age at diagnosis was 66 years (range, 23-88 years). There were seven patients who had pathology and demographics data available without any further clinical information. Of the remaining 58 patients, 48% presented with resectable/borderline-resectable disease (n = 29), 9% with locally advanced/unresectable disease (n = 5), and 43% were metastatic disease (n = 25) at diagnosis (Table 1).

TABLE 1.

Clinical and Pathologic Characteristics of Patients With Pancreatic Acinar Cell Carcinoma

Characteristic	Overall (n = 65)
Sex, no. (%)
Male	48 (74)
Female	17 (26)
Age at diagnosis, years
Median (range)	66 (23-88)
Ethnicity, no. (%)
White	47 (87)
Black	3 (6)
Others	4 (7)
Tumor location in the pancreas, no. (%)
Head	25 (46)
Body/tail	29 (54)
Unknown (imaging absent)	11
Tumor size at presentation, cm
Mean (SD)	5.4 (3.5)
Stage at diagnosis, no. (%)
Resectable/borderline-resectable	29 (48)
Locally advanced	5 (9)
Metastatic	25 (43)
Histology, no. (%)
Pure acinar	51 (79)
Mixed acinar-neuroendocrine	11 (17)
Mixed acinar-adenocarcinoma	1 (2)

Abbreviation: SD, standard deviation.

Nearly all patients (93%, n = 27/29) with resectable/borderline-resectable disease underwent surgical resection (Table 2). The two patients without resection had underlying comorbidities that precluded surgical intervention. Over half (59%, n = 17/29) of these patients received perioperative chemotherapy with folinic acid, fluorouracil, irinotecan, and oxaliplatin (FOLFIRINOX) as the most common regimen in both the neoadjuvant (n = 11) and adjuvant (n = 3) settings. Perioperative radiation therapy was also a common treatment option (41%, n = 12/29), with most of the patients receiving treatment of the primary tumor in the neoadjuvant setting with concurrent chemotherapy. One patient received neoadjuvant radiation alone after declining concurrent chemotherapy. The objective response rate (ORR) for patients who received induction chemotherapy plus perioperative radiation was 73% (n = 8/11; 81% n = 8/9 for neoadjuvant chemotherapy plus chemoradiation, 0% n = 0/2 for neoadjuvant chemotherapy plus adjuvant chemoradiation) versus 66% (n = 2/3) for neoadjuvant induction chemotherapy alone.

TABLE 2.

Perioperative Treatment Summary and Surgical Outcomes for Patients With Resectable/Borderline-Resectable Pancreatic Acinar Cell Carcinoma

Characteristic	Overall (n = 29), no. (%)
Perioperative chemotherapy	17 (59)
Neoadjuvant chemotherapy	15 (88)
FOLFIRINOX	11 (73)
Gemcitabine-based	2 (13)
Platinum + etoposide	2 (13)
Adjuvant chemotherapy	4 (24)
FOLFIRINOX	3 (75)
Gemcitabine-based	1 (25)
Perioperative radiation	12 (41)
Neoadjuvant	10 (83)
Adjuvant	2 (17)
Resection	27 (93)
Whipple	11 (41)
Distal pancreatectomy	13 (48)
Total pancreatectomy	3 (11)
Node positive	14 (52)
Lymphovascular invasion	9 (33)
Margin positive	5 (19)
Recurrence	9 (31)
Liver	5 (56)
Pancreas	2 (22)
Lymph nodes	2 (22)

Abbreviation: FOLFIRINOX, folinic acid, fluorouracil, irinotecan, and oxaliplatin.

Of the 27 patients who underwent surgical resection, one third of the patients (33%, n = 9) had known recurrence, with five patients lost to follow-up. In the recurrent population, majority had pure acinar cell carcinoma (n = 7), followed by acinar-neuroendocrine (n = 2), with primary tumor location in the head of the pancreas (78%, n = 7/9). Three had margin-positive resection (33%), while the rest had node-positive disease (66%, n = 6/9) or lymphovascular invasion (33%, n = 3/9). Patients who received neoadjuvant FOLFIRINOX had a recurrence rate of 36% (n = 4). Both patients who received cisplatin plus etoposide (100%, n = 2) in the neoadjuvant setting had recurrence. Furthermore, both patients who received adjuvant radiation to the tumor bed had recurrence, while only one third of the patients who received concurrent neoadjuvant chemoradiation (33%, n = 3) had recurrence. Six patients (66%) with recurrence had somatic NGS testing, most of whom (83%, n = 5/6) had pathogenic alterations identified (FGFR1 amplification, BRCA2, NTRK1 fusion, MTOR, and BRAF V600E, respectively). Six patients (66%) had germline NGS testing, none with pathogenic mutations, and 1 with a variant of uncertain significance (VUS) in the MLH3 gene. The median time to follow-up in the resectable population was 40 months, with 3-year and 5-year survival rates of 79% and 72%, respectively. The median recurrence-free survival (RFS) for patients who underwent curative intent with surgical resection was 36 months with a 47% RFS for both the 3-year and 5-year marks (Fig 1).

FIG 1.

Kaplan-Meier survival curves for RFS (A) and OS (B) in patients with resectable/borderline-resectable pancreatic acinar cell carcinoma. ^aKaplan-Meier method. Est, estimated; KM, Kaplan-Meier; NE, not estimable; OS, overall survival; RFS, recurrence-free survival.

Most of the patients in the recurrent population had metachronous metastases to the liver (56%, n = 5/9), followed by local recurrence in the pancreas (22%, n = 2/9) and lymph nodes only (22%, n = 2/9; Table 2). Median RFS for patients with liver metastases was 7.2 months compared with 10.7 months and 17.0 months for patients with isolated lymph node involvement and local recurrence, respectively. There were no obvious differences in genomic alterations with respect to the location of disease recurrence. Three of the five patients with metachronous metastases in the liver had pathogenic mutations (FGFR1 amplification, BRCA2, and MTOR). The patients with the NTRK fusion and BRAF V600E mutation recurred within the pancreas and lymph nodes, respectively.

There were eight patients (29.6%) who underwent resection and had no recurrence at the 5-year mark. Most were pure acinar cell carcinoma (88%, n = 7/8) with primary tumor location in the body/tail of the pancreas (75%, n = 6/8). Majority were node negative (75%, n = 6/8) without lymphovascular invasion (88%, n = 7/8). Half (50%, n = 4/8) received neoadjuvant FOLFIRINOX and radiation. There was 1 patient who received neoadjuvant gemcitabine plus cisplatin without radiation, while the remaining patients (n = 3) had neither chemotherapy nor radiation in the perioperative period. None of these patients had somatic NGS testing. Two patients had germline testing, one of whom had a germline BRCA2 mutation and the other had a VUS in the MLH1 gene. In patients with locally advanced/metastatic disease, only 1 patient underwent surgical resection for oligometastatic disease to the liver. He received neoadjuvant folinic acid, fluorouracil, and oxaliplatin (FOLFOX) for 4 months before resection with partial radiographic response to the primary pancreatic mass and liver metastasis. He had progressive disease to the liver approximately 3 years after surgery and was treated with gemcitabine plus cisplatin for 4 months before being lost to follow-up. Most (95%, n = 21/22) of the patients with locally advanced/metastatic disease received palliative systemic therapy. The remaining patient was placed on hospice before beginning any systemic therapy. First-line treatments varied, with 5-FU–based chemotherapy accounting for majority of the selected regimens (n = 10 FOLFIRINOX, n = 1 folinic acid, fluorouracil, and irinotecan [FOLFIRI]), followed by gemcitabine-based chemotherapy (n = 2 gemcitabine plus nab-paclitaxel, n = 1 gemcitabine plus cisplatin, n = 1 gemcitabine plus erlotinib), and carboplatin plus etoposide (n = 3). Other first-line treatments included topotecan (n = 1) and capecitabine plus temozolomide (n = 1), both of which progressed after only 1 month. With a median follow-up of 20.4 months, median progression-free survival (PFS) and OS of patients with locally advanced/metastatic PACC in the first-line setting were 10.8 and 27.3 months, respectively. When analyzed separately, median OS for patients with locally advanced and metastatic disease was 42.5 months and 24.7 months, respectively (Fig 2).

FIG 2.

Kaplan-Meier survival curves for PFS (A) and OS (B) in patients with locally advanced/metastatic pancreatic acinar cell carcinoma. ^aKaplan-Meier method. Est, estimated; KM, Kaplan-Meier; OS, overall survival; PFS, progression-free survival.

There were 68% (n = 15) of patients who received second-line treatment, most of which were 5-FU based (n = 4 FOLFIRINOX, n = 2 FOLFIRI, n = 1 FOLFOX), followed by gemcitabine-based chemotherapy (n = 2 gemcitabine plus nab-paclitaxel, n = 1 gemcitabine plus cisplatin, n = 1 gemcitabine plus capecitabine). Descriptive outcomes of patients across different lines of therapy including disease control rate (DCR), ORR, and PFS are summarized in Table 3.

TABLE 3.

Descriptive Outcomes for Patients With Locally Advanced/Metastatic Pancreatic Acinar Cell Carcinoma According to Line of Therapy

First-Line Therapy	Total (n = 23)	5-FU Based (n = 12)	Gem Based (n = 5)	Platinum Based (n = 4)	Others (n = 2)
Progression, no. (%)	6 (26)	3 (25)	0	2 (50)	1 (50)
DCR, no. (%)	15 (65)	8 (67)	4 (80)	2 (50)	1 (50)
ORR, no. (%)	10 (43)	7 (58)	2 (40)	0	0
PFS, months	11.8 (9.0-NE)	11.8 (11.1-NE)	NA	5.3 (2.4-NE)	9.0 (NE-NE)
Second-Line Therapy	Total (n = 17)	5-FU Based (n = 8)	Gem Based (n = 5)	Platinum Based (n = 0)	Others (n = 4)
Progression, no. (%)	10 (59)	5 (63)	3 (60)	0	2 (50)
DCR, no. (%)	7 (35)	3 (38)	2 (40)	0	1 (25)
ORR, no. (%)	3 (19)	1 (13)	2 (40)	0	0
PFS, months	11.4 (9.0-NE)	11.8 (11.1-NE)	9.0 (8.2-NE)	NA	NA
Third-Line Therapy	Total (n = 11)	5-FU Based (n = 3)	Gem Based (n = 4)	Platinum Based (n = 0)	Others (n = 4)
Progression, no. (%)	7 (64)	2 (67)	2 (50)	0	3 (75)
DCR, no. (%)	3 (27)	1 (33)	1 (25)	0	1 (25)
ORR, no. (%)	0	0	0	0	0
PFS, months	11.1 (7.0-NE)	9.0 (8.2-NE)	11.8 (11.1-NE)	NA	5.3 (2.4-NE)

Abbreviations: 5-FU based, FOLFIRINOX, FOLFIRI, or FOLFOX; DCR, disease control rate (stable, partial, or complete response); FOLFIRI, folinic acid, fluorouracil, and irinotecan; FOLFIRINOX, folinic acid, fluorouracil, irinotecan, and oxaliplatin; FOLFOX, folinic acid, fluorouracil, and oxaliplatin; Gem based, gemcitabine plus nab-paclitaxel, gemcitabine plus cisplatin, gemcitabine plus ertolitinib, or gemcitabine plus oxaliplatin; NA, not available; NE, not estimable; ORR, objective response rate (partial or complete response); PFS, progression-free survival; Platinum based, carboplatin plus etoposide or cisplatin plus etoposide.

Somatic NGS Testing

Somatic NGS and molecular testing were performed in 21 patients (Fig 3) and led to additional therapeutic options in 43% (n = 9/21) of patients. Notably, several patients had concomitant alterations in homologous recombination (HR), mismatch repair (MMR), and BRAF genes, which allowed for the use of multiple targeted agents for two patients. One patient received ipilimumab and nivolumab for mismatch repair deficient (dMMR) PACC (MLH1 and PMS2 loss by IHC staining) and eventually olaparib for HR mutations (PALB2, RAD51C). Another patient received olaparib for a germline HR mutation (CHEK2) and eventually dabrafenib plus trametinib for BRAF alteration (SND1-BRAF fusion). Most of the patients had pathogenic alterations in the HR pathway (39%, n = 9) with ATM (n = 4), BARD1 (n = 1), BRCA2 (n = 4), CHEK2 (n = 1), PALB2 (n = 3), and RAD51C (n = 1). Three of these patients received targeted therapy with olaparib during their disease course. One patient was started on maintenance olaparib after achieving complete radiographic response to chemotherapy (FOLFIRINOX ×6 months, FOLFIRI ×3 months) and has sustained disease control for over 17 months. The two other patients experienced progressive disease after only 1-2 months.

FIG 3.

Pathogenic mutations identified by somatic NGS (A) and germline NGS (B) for patients with pancreatic acinar cell carcinoma. HRR, homologous recombination repair; MMR, mismatch repair; NGS, next-generation sequencing; VUS, variant of uncertain significance.

The next most common mutations identified in patients with somatic NGS testing were RAF alterations (26%, n = 6). These included BRAF V600E (n = 1), KANK4-RAF1 fusion (n = 1), SND1-BRAF fusion (n = 2), TBXAS1-BRAF fusion (n = 1), and BRAF intron 9 inversion (n = 1). Three of these patients had targeted therapy with BRAF inhibitors (n = 2 dabrafenib plus trametinib, n = 1 ulixertinib plus trametinib). Two patients progressed after only 1 month of treatment, and the other was lost to follow-up.

Three patients (13%) had mutations in the MMR genes. One patient with TBXAS1-BRAF rearrangement on NGS was also found to have MLH1 and PMS2 loss by IHC staining of the primary tumor. He received combination immunotherapy with nivolumab plus ipilimumab. Unfortunately, there was progression after only 1 month of immunotherapy. Two other patients had MMR mutations identified on NGS testing: one patient with MLH1 and MSH3 without confirmatory IHC staining and another patient with MLH1 who was MMR proficient on IHC. However, neither patient received immunotherapy because of stable disease.

There were two other patients who were given targeted therapy over their disease course, including 1 patient with ETV6-NTRK3 gene fusion who received treatment with larotrectinib (tropomyosin receptor kinase inhibitor) in the first line. He received 18 months of therapy with partial radiographic response before discontinuing the medication as a result of side effects of anasarca and anemia. The other patient was found to have a FGFR1 mutation and received treatment with sunitinib in the second line after FOLFIRINOX. He sustained stable disease control over 20 months before progression, when he was transitioned to gemcitabine plus nab-paclitaxel.

Germline NGS Testing

Germline NGS and molecular testing were performed in 21 patients, with 43% (n = 9/21) having pathogenic mutations and 10% (n = 2/21) with VUS. Similar to somatic NGS results, most of the patients (33%, n = 7) had alterations in the HR pathway, including mutations to ATM, CHEK2, BARD1, PALB2, and BRCA2. Two of the patients with somatic HR mutations who were treated with olaparib also had germline alteration in HR genes (ATM and PALB2). None of the patients were found to have germline alterations in MMR genes; however, two patients were noted to have VUS to MLH1 and MSH3, respectively. One of these patients was MMR proficient on IHC, whereas the other had no IHC testing completed. Neither received immunotherapy. Finally, there was 1 patient with a germline VUS to the BLM and CDKN1B genes. The patient's somatic NGS identified a KANK4-RAF1 fusion described previously and eventually received treatment with dabrafenib plus trametinib.

DISCUSSION

The clinical characteristics of PACC from our study are consistent with prior reports with earlier-in-life presentation and the higher incidence in male patients compared with the clinical characteristics of PDAC.^5,13,14 The pathologic features of PACC are also distinct, with occasional ductal and neuroendocrine differentiation.^10,12,15 A similar case series of 43 patients with PACC showed six patients (14%) with neuroendocrine features, which is consistent with the 17% identified in our cohort.¹⁶ In addition, we found that almost half of these patients presented with resectable/borderline-resectable disease, which is also similar to other multi-institution studies showing earlier stages (I and II) accounting for 45%-55% and later stages (III and IV) between 40% and 50% at diagnosis.^5,17,18 Among the various clinical outcomes measured was the R0 resection rate at 81% (Table 2). A separate retrospective study of patients with PACC in the National Cancer Database (NCDB) showed similar outcomes with a R0 resection rate of 77% compared with 72% in a similar review for patients with PDAC.^19,20 Patients in our study who underwent surgical resection had a median RFS of 36 months, which is better than other studies reported in the literature.²¹ However, most of these studies report smaller sample sizes and used older regimens. A recent retrospective multicentric study showed median RFS of 30 months and 5-year RFS of 38%, which is comparable with our results.²² The difference in clinical outcomes is likely because of the rarity of the disease and variable patient populations between studies.

Despite the discrepancy in clinical outcomes between studies, PACC in general has more favorable prognosis compared with PDAC, particularly in patients with resectable/borderline-resectable disease.^{4,12,19,23,24} Prior studies have shown conflicting results regarding the benefit of adjuvant chemotherapy in PACC.^18,22 However, given the small sample size in such studies, it is hard to draw definitive conclusions. Therefore, large database studies could be useful in assessing such a question. A recent NCDB analysis showed that adjuvant therapy was associated with survival benefit compared with resection alone.²⁵ Although we did not perform such analysis in our study given the small sample size, it is important to note that most (n = 5/8) of the patients who were alive and free of disease at 5 years received neoadjuvant platinum-based therapy in the perioperative setting. Given that the role of perioperative chemotherapy is well established in PDAC, it is reasonable to extrapolate that neoadjuvant strategies would be of value in PACC, especially given the higher rates of HR gene alterations in these patients and the sensitivity of such tumors to platinum agents.²⁶ Even as DCR and ORR declined with successive lines of therapy in patients with advanced/metastatic disease, median PFS remained consistent across all treatment lines, ranging between 11 and 12 months (Table 3). There is limited research directly comparing treatment approaches within this patient demographic; however, findings from other retrospective cohorts align closely with our own results.^7,17,27

PACC is a genomically diverse malignancy that is manifested by the significant number of patients with somatic and germline mutations in our cohort.²⁸ Consistent with prior studies, most common alterations were in the HR pathway and RAF/BRAF.^29-33 These somatic changes are in contrast to what is seen in patients with PDAC where KRAS and TP53 are more common.^31,34 A recently published study compared whole-exome sequencing and NGS of pure PACC specimens with a large array of PDAC samples.³¹ They showed genomic loss of heterozygosity in patients with PACC was frequently associated with alterations to HR, including BRCA1/BRCA2, and PALB2. Conversely, mutations to TP53, CDKN2A, and KRAS were more likely to be found in patients with PDAC. The study also identified BRAF V600E mutation but only two patients with PACC (3.8%) without other RAF alterations, a finding the authors speculate may be related to the inclusion of only pure PACC patient cases.

In addition to somatic mutations, we found germline mutations in 43% of the patients with PACC, with the majority in the HR pathway (33%). This relatively high incidence is consistent with a recent study of 49 patients with PACC, which found HR alterations in 37% of the patients.³⁵ This emphasizes the importance of universal germline testing for all patients with PACC, consistent with current guideline recommendations for PDAC.³⁶

NGS testing led to therapeutic implications in 43% (n = 9/21) of the patients in our study with BRAF/MEK inhibitors, checkpoint inhibitors, PARP inhibitors, and the TRK inhibitor. Therefore, we emphasize the importance of checking for these alterations, as there are US Food and Drug Administration (FDA)–approved, tissue-agnostic indications for solid tumors: larotrectinib for NTRK gene fusions, dabrafenib and trametinib for BRAF-mutated tumors, and pembrolizumab or dostarlimab for dMMR.^37-40 Although there is currently no tissue-agnostic indication for HR alterations, identifying such tumors is important in systemic treatment selection with their sensitivity to platinum agents.

This study has multiple limitations, most notably the retrospective design and small sample size, which precluded meaningful comparisons between specific chemotherapy regimens or treatment strategies. In addition, the somatic and germline NGS testing methods varied as the results were used from commercially available assays. However, we believe that this study represents an important addition to the PACC literature in terms of describing the clinical characteristics and diverse molecular features of this rare entity.

In conclusion, PACC is a rare malignancy that requires further study to determine optimal treatment strategies. Currently, a multidisciplinary approach is integral in treating these patients especially in the earlier stages. In addition, NGS testing should be performed as a significant number of patients have tumors that are enriched with somatic alterations that can potentially lead to FDA-approved indications. As recommended by the guidelines for pancreatic cancer, universal germline testing is recommended in patients with PACC given the high prevalence of germline mutations in this population.

AUTHOR CONTRIBUTIONS

Conception and design: Cody Eslinger, Thorvardur R. Halfdanarson, Mohamad Bassam Sonbol

Financial support: Mohamad Bassam Sonbol

Provision of study materials or patients: Rish Pai

Collection and assembly of data: Cody Eslinger, Bobak Seddighzadeh, Zaid Elsabbagh, Chris Hartley, Thorvardur R. Halfdanarson, Mohamad Bassam Sonbol

Data analysis and interpretation: Cody Eslinger, Claire Yee, Rish Pai, Jason Starr, Tanios Bekaii-Saab, Thorvardur R. Halfdanarson, Mohamad Bassam Sonbol

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/po/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).