Clinical characterization of patients with gBRCA1/2 mutation-positive unresectable pancreatic cancer: a multicenter prospective study

Abstract

Background

Accumulating evidence has demonstrated platinum-based chemotherapy followed by maintenance therapy with a poly Adenosine diphosphate (ADP)-ribose polymerase inhibitor (olaparib) show benefits in unresectable pancreatic cancer with a germline (g)BRCA1/2 mutation. Evaluation of the germline BRCA1 and BRCA2 mutation is essential for making decisions on a treatment strategy for patients with unresectable pancreatic cancer. However, the detection rates of germline BRCA1 and BRCA2 mutations and efficacy of maintenance with olaparib remain undetermined, prospectively, in Japan.

Methods & Results

In this prospective analysis, the rate of germline BRCA1 and BRCA2 mutations and efficacy of chemotherapy were analyzed in 136 patients with pancreatic cancer who underwent BRACAnalysis® (85 patients) or FoundationOne® CDx (51 patients) between January 2020 and July 2022. A total of six patients (4.4%) had a germline BRCA1 and BRCA2 mutation. Five patients were treated with modified FOLFIRINOX and one with fluorouracil and oxaliplatin. All patients continued platinum-based chemotherapy for ˃4 months and were subsequently treated with olaparib as a maintenance therapy. The response rate to platinum-based chemotherapy in the germline BRCA1 and BRCA2 mutation-positive group was significantly better than that of the germline BRCA1 and BRCA2 mutation-negative group (66% vs 23%, P = 0.04). All patients harbouring a germline BRCA1 and BRCA2 mutation were able to switch to olaparib. The median progression-free survival using olaparib was 5.7 months (range 3.0–9.2).

Conclusions

The rate of germline BRCA1 and BRCA2 mutations found in patients with unresectable pancreatic cancer was comparable to those of previous studies.

An analysis of germline BRCA1 and BRCA2 mutations has benefits for all patients with unresectable pancreatic cancer with regard to decisions on therapeutic strategies in a clinical practice setting.

An analysis of gBRCA1/2 mutations has benefits for all patients with unresectable pancreatic cancer with regard to decisions on therapeutic strategies in a clinical practice setting.

Introduction

Pancreatic cancer remains a disease with a poor prognosis, accounting for almost as many deaths as cases in the world in 2020 (1). Only in 10–20% of patients with pancreatic cancer can undergo a curative resection owing to locally advanced or metastatic disease (2). However, even after a resection, over 75% will experience a recurrence within 5 years (3). Chemotherapy is often administered to patients with locally advanced, metastatic disease and recurrence. Recent data suggests that treatment with a combination regimen of 5-fluorouracil, leucovorin, oxaliplatin and irinotecan (FOLFIRINOX) (4), and a regimen of gemcitabine plus nab-paclitaxel (GnP) (5) has improved the prognosis of pancreatic cancer patients with metastatic disease, but the median overall survival (OS) is still ˂1 year. Recently, attempts to identify a population of patients with pancreatic cancer who would benefit from targeted therapy have led to the selection of a cohort showing germline (g)BRCA1 and BRCA2 (BRCA1/2) mutations. According to a previous report, from 4 to 7% of patients with pancreatic cancer have a germline BRCA1 and BRCA2 (gBRCA1/2) mutation (6–9). BRCA1/2 are genes that code for tumour suppressor proteins involved in the repair of double-stranded DNA breaks via homologous recombination deficiency (10). Mutations in both genes are associated with the risk of development of several cancer types, including breast, ovarian, prostate, oesophageal, gastric, biliary tract and pancreatic cancer (8). Patients with gBRCA1/2 mutations have a better prognosis and from 67.4% to 76.3% of patients achieved a good response to platinum-based chemotherapy (11,12). This is due to the inability of BRCA mutation tumour cells to repair double-stranded DNA breaks, which require homologous recombination to repair, generated by platinum-based compounds (13). Also, poly ADP-ribose polymerase (PARP) enzymes primarily repair single-stranded DNA breaks and play crucial roles in DNA damage repair (DDR). Accumulation of single-stranded DNA breaks in the presence of PARP inhibitors results in the formation of double-stranded breaks. As mentioned above, BRCA mutation tumour cells are unable to repair double-stranded DNA, and accumulate such DNA breaks over time, resulting in cell death (13). The POLO trial revealed maintenance olaparib after platinum therapy prolonged progression-free survival (PFS) (14). Patients with a gBRCA1/2 mutation may have a prolonged prognosis with PARP inhibitor treatment. Therefore, the evaluation of gBRCA1/2 mutations is essential for decisions on treatment strategies for patients with unresectable pancreatic cancer. However, the detection rates for gBRCA1/2 mutations and efficacy of maintenance with olaparib are undetermined.

In this study, we aimed to elucidate the characteristics of gBRCA1/2 mutations in unresectable pancreatic cancer and the efficacy of chemotherapy for gBRCA1/2 mutation-positive pancreatic cancer.

Patients and methods

Patient identification and data collection

This study is a multicenter, prospective observational study of the detection of gBRCA1/2 mutations in patients with postoperative recurrence, locally advanced unresectable (UR-LA) and metastatic unresectable (UR-M) pancreatic cancer. Patients were eligible if they were 18 years of age or older so that they could give consent, and had histologically or radiologically confirmed pancreatic cancer. Patients for whom chemotherapy was not appropriate were excluded. Resectability was defined using National Comprehensive Cancer Network (NCCN) guidelines v1.2019 (15). All relevant institutional ethics review boards approved this study (332–127), which was performed according to the provisions of the Declaration of Helsinki. Written consent was obtained from all patients. One hundred and thirty-six patients with pancreatic cancer underwent gBRCA1/2 analysis using BRACAnalysis® (Myriad Genetics Laboratories, Inc., Salt Lake, UT, USA) or FoundationOne® CDx (Foundation Medicine, Inc., Cambridge, MA, USA). Patients were determined to be gBRCA1/2 mutation-positive based on an algorithm (Supplementary Fig. 1). Since the diagnosis of germlines in FoundationOne® CDx is not possible, single-site sequencing of mononuclear cells was performed in cases with a positive BRCA1/2 mutation according to FoundationOne® CDx.

Patients were recruited between 1 January 2020 and 31 July 2022. Clinical data on patient demographics, clinical characteristics and response to treatment were collected from medical records. Systemic chemotherapy was administered as a standard of care therapy in terms of dosing and schedule. Dose modification as a result of toxicity was conducted at the treating physician’s discretion. Responses to treatment were analyzed using Response Evaluation Criteria in Solid Tumours v1.1 (16). Adverse events were graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events, version 4.0.

Statistical analysis

Patient demographics, clinical characteristics and response to treatment were calculated and stratified according to being gBRCA1/2 mutation-positive or -negative using Fisher’s exact test. The PFS was calculated from the time of initiation of olaparib until evidence of progression, death or last follow-up. The OS was calculated from the date of a diagnosis as an unresectable cancer and initiation of chemotherapy until death. The time to treatment failure (TTF) observation period was set at 4 months in platinum-based chemotherapy because gBRCA1/2 mutation-positive patients were switched to using a PARP inhibitor after 4 months on such chemotherapy. Data were finalized on 31 March 2023; all patients still alive and not showing disease progression were censored at that time. P values were two-sided, and considered statistically significant when ˂0.05. All analyses were conducted using EZR 1.55 (17) and GraphPad Prism 9.1.1(GraphPad Software, San Diego, CA, USA).

Results

Patient characteristics and the rate of gBRCA1/2 mutations in unresectable pancreatic cancer

Overall, 136 patients with pancreatic cancer underwent BRCA genetic analysis. The median age at diagnosis was 69 years (range 37–88 years). Seventy patients (51.4%) were male and 109 patients (80.1%) were Eastern Cooperative Oncology Group Performance Status 0 (Supplementary Table 1). The proportions of each of the anatomic locations and pathological diagnosis were comparable to those previously reported (3). Forty-three patients (31.6%) had recurrent disease, and 93 (68.3%) had unresectable disease. As first-line chemotherapy, GnP and platinum-based chemotherapy [modified FOLFIRINOX (mFFX) or FOLFOX] were administered to 61 (44.8%) and 54 (39.7%) patients, respectively. Twenty-four patients (17.6%) underwent BRCA genetic analysis prior to treatment, 50 patients (36.8%) during first-line treatment, and 62 patients (45.6%) after the start of second-line treatment. BRCA genetic analysis was undertaken by 85 patients (62.5%) using BRACAnalysis® and by 51 patients (37.5%) using FoundationOne® CDx. Six patients (4.4%) had a gBRCA1/2 deleterious mutation (BRCA1: two patients, BRCA2: four patients) and six patients (4.4%) had a BRCA1/2 variant of uncertain significance.

Demographic and clinical characteristics of patients with gBRCA1/2 mutation-positive or -negative pancreatic cancer

No significant difference was noted in age, sex, anatomic location and pathological diagnosis between gBRCA1/2 mutation-positive and -negative groups was noted. Of note, a family history of cancer was defined as having a first, second or third-degree relative with pancreatic, breast, ovarian or prostate cancer. A significant difference in a family history of cancer was observed between the two groups (Table 1). Thirty patients (23.1%) of the gBRCA1/2 mutation-negative group had a positive family history of cancer, whilst four patients (66.7%) of the gBRCA1/2 mutation-positive group had a positive family history of cancer (P = 0.034, hazards ratio = 6.54). The family history for each type of cancer is as presented in Supplementary Fig. 2. Regarding pathological diagnosis, five (5/114, 4.4%) patients with adenocarcinoma had a gBRCA1/2 mutation and one (1/2; 50%) patient with acinar cell carcinoma had a gBRCA2 mutation; however, these differences were not statistically significant (P = 0.12).

Table 1.

Demographic and clinical characteristics of gBRCA1/2 mutation-positive and -negative pancreatic cancer patients

		gBRCA1/2 mutation-positive (n = 6)	gBRCA1/2 mutation-negative (n = 130)	P value
Age				1.000
	Years (median)	66 (59–77)	69 (37–88)
	Age at diagnosis (30–50 years)	0	8 (6.2)
	Age at diagnosis (51–90 years)	6 (100)	122 (93.8)
Sex				1.000
	Male	3 (50)	67 (51.5)
	Female	3 (50)	63 (48.5)
Family history				0.034
	Negative	2 (33.3)	100 (76.9)
	Positive	4 (66.7)	30 (23.1)
Anatomic location				0.119
	Head	3 (50.0)	67 (51.5)
	Body	2 (33.3)	40 (30.8)
	Tail	0	22 (16.9)
	Others	1 (16.7)	1 (0.8)
Pathological diagnosis
	Adenocarcinoma	5 (83.3)	109 (83.8)	0.126
	Acinar cell carcinoma	1 (16.7)	1 (0.8)
	Others	0	20 (15.4)

Family history: 2nd-degree relatives.

Efficacy of platinum-based chemotherapy in patients with gBRCA1/2 mutation-positive or -negative pancreatic cancer

Six gBRCA1/2 mutation-positive patients (100%) and 76 gBRCA1/2 mutation-negative patients (58.4%) were administered platinum-based chemotherapy. Half of all gBRCA1/2 mutation-positive patients were administered platinum-based chemotherapy as second-line treatment, although there was no significant difference in treatment line compared to gBRCA1/2 mutation negative-patients (Table 2). Platinum-based chemotherapies were mostly modified FFX in both groups. For gBRCA1/2 mutation-positive patients, four (66.7%) showed a partial response (PR), and two (33.3%) showed stable disease (SD). All gBRCA1/2 mutation-positive patients continued with platinum-based chemotherapy for 4 months. In contrast, 39.7% of gBRCA1/2 mutation-negative patients failed to continue with platinum-based chemotherapy by 4 months (Fig. 1A). Platinum-based chemotherapy resulted in tumour shrinkage in all gBRCA1/2 mutation-positive patients (Fig. 1B). The disease control rate (DCR) was numerically higher in the gBRCA1/2 mutation–positive group (100% vs. 56.6%; P = 0.076). The overall response rate (ORR) was significantly higher for gBRCA1/2 mutation-positive patients compared to gBRCA1/2 mutation-negative patients (66.7% vs. 23.7%; P = 0.041).

Table 2.

Details of a platinum regimen in gBRCA1/2 mutation-positive and -negative pancreatic cancer patients

		gBRCA1/2 mutation-positive (n = 6)	gBRCA1/2 mutation-negative (n = 76)	P value
Treatment line				0.330
	1st	3 (50)	54 (71.1)
	2nd	3 (50)	17 (22.4)
	3rd	0	5 (6.6)
Platinum based regimen				0.376
	Modified FFX	5 (83.3)	71 (93.4)
	FOLFOX	1 (16.7)	5 (6.6)
Best overall response				0.115
(RECIST v1.1)
	CR	0	0
	PR	4 (66.7)	18 (23.7)
	SD	2 (33.3)	25 (32.9)
	PD	0	18 (23.7)
	NE	0	15 (19.7)
	DCR	100%	56.60%	0.076
	ORR	66.70%	23.70%	0.041

CR: complete response; DCR: disease control rate; FFX: fluorouracil, irinotecan, and oxaliplatin; FOLFOX: fluorouracil and oxaliplatin; NE: not evaluated; ORR: overall response rate; PD: progressive disease; PR: partial response; SD: stable disease; RECIST: Response Evaluation Criteria in Solid Tumours.

Figure 1.

Efficacies of platinum-based chemotherapy. (A) Kaplan–Meier estimates of time to treatment failure of platinum-based chemotherapy in gBRCA1/2 mutation-positive and -negative groups. (B) Change in tumour size with platinum-based chemotherapy in gBRCA1/2 mutation–positive patients. g: germline; PR: partial response; SD: stable disease; TTF: time to treatment failure; VUS: variant of uncertain significance.

Treatment details and efficacy of olaparib in patients with gBRCA1/2 mutation-positive pancreatic cancer

The median age of gBRCA1/2 mutation-positive patients was 66 years (range 59–77). Only one of the six cases had a history of cancer, this being breast cancer. None of the patients had a known gBRCA1/2 mutation prior to the diagnosis of pancreatic cancer. Two (33.3%) patients had a gBRCA1 mutation, and four (66.6%) patients had a gBRCA2 mutation. Two patients underwent assessment using FoundationOne® CDx and four patients underwent assessment using BRACAnalysis® (Table 3). Case 1 was diagnosed with pancreatic cancer before insurance coverage became available for FoundationOne® CDx and olaparib. Case 1 had already been treated with platinum-based chemotherapy for ˃1 year when olaparib became available. Other cases underwent BRACAnalysis® immediately after the diagnosis of pancreatic cancer and were switched to olaparib after 4–7 months of platinum-based chemotherapy. Some cases required a reduction in the platinum regimen due to treatment-related adverse events (TRAE), and one case was switched to olaparib because they developed an intolerance to platinum-based chemotherapy. The other cases were switched to olaparib aggressively at the direction of the treating physician when their tumour shrank during platinum-based chemotherapy. Accordingly, all gBRCA1/2 mutation-positive patients were treated with olaparib (Fig. 2A). Regarding TRAE, Grade(Gr)2 nausea and fatigue and Gr1 fatigue were observed in one patient each. All cases started at 600 mg/day; five out of six cases continued at the same dose until the last dose, and one case was reduced to 500 mg/day due to Gr2 fatigue. None of the patients had to discontinue olaparib due to TRAE. Regarding the best overall response to olaparib, one and three patients showed PR and had SD, respectively. Thus, the ORR was 16.7% (Supplementary Table 2). The median progression-free survival of gBRCA1/2 mutation-positive patients was 5.7 months (range 3.0–9.2; Fig. 2B). The median OS was 46.6 months [95% confidence interval (CI), 32.1–61.2] in the gBRCA1/2 mutation-positive group compared with 21.7 months (95% CI, 19.5 to 23.9) in the gBRCA1/2 mutation-negative group (P = 0.354; Supplementary Fig. 3). During the observation period, four out of six cases were judged as progressive disease (PD) and subsequently olaparib was discontinued. Amongst PD cases, three and one were treated with platinum-based chemotherapy and GnP, respectively. Two out of three cases treated with platinum-based chemotherapy after olaparib failure were judged as PD at 1.5–2 months from the re-initiation of a platinum-based regimen, suggesting that a platinum regimen may not be effective after olaparib failure.

Table 3.

Details of gBRCA1/2 mutation-positive pancreatic cancer patients

No.	Age	Sex	Tumour status	Pathologic diagnosis	Gene	Variant	Timing of genomic analysis	Method of examination	Regimen of platinum-based chemotherapy
1	59	M	Recurrence	Acinar cell carcinoma	BRCA1	c.188 T > A	During 2nd line therapy	FoundationOne®	mFOLFIRINOX
2	62	F	UR-M	Adenocarcinoma	BRCA2	c.1278del (p.Asp427Thrfs*3) Heterozygous	Pretreatment	BRACAnalysis®	mFOLFIRINOX
3	68	F	UR-LA	Adenocarcinoma	BRCA1	c.2269del(p.Val757Phefs*8) Heterozygous	During 1st line therapy	BRACAnalysis®	mFOLFIRINOX
4	63	M	Recurrence	Adenocarcinoma	BRCA2	frameshift N1287fs*1	During 3rd line therapy	FoundationOne®	mFOLFIRINOX
5	77	F	UR-M	Adenocarcinoma	BRCA2	c.6952C > T(p.Arg2318*)	During 1st line therapy	BRACAnalysis®	FOLFOX
6	69	M	UR-LA	Adenocarcinoma	BRCA2	c.6952C > T(p.Arg2318*)	During 1st line therapy	BRACAnalysis®	mFOLFIRINOX

F: female; M: male; UR-LA: unresectable locally advanced; UR-M: unresectable metastatic.

Figure 2.

Clinical course and progression-free survival in gBRCA1/2 mutation-positive patients after olaparib therapy. (A) Clinical course. (B) Kaplan–Meier estimates of progression-free survival of gBRCA1/2 mutation-positive patients after olaparib therapy. FFX: fluorouracil, irinotecan, and oxaliplatin; FOLFOX: fluorouracil and oxaliplatin; GnP: gemcitabine and nab-paclitaxel; mFFX: modified FFX; g: germline; PD: progressive disease; PFS: progression-free survival; PR: partial response; SD: stable disease.

Genetic analysis of family members of gBRCA1/2 mutation-positive patients

Four out of six families underwent genetic counselling. Of the two families that did not undergo genetic counselling, one family refused and the other family’s details are unknown. Three families underwent genetic testing based on the patient’s genetic data. The son and daughter of case1 underwent genetic counselling and testing. The same gBRCA1 mutation was detected in the son and his surveillance was initiated. The sister of case 2, and brother and son of case 4 underwent genetic counselling and testing and no gBRCA mutations ware found (Supplementary Table 3).

Discussion

In this study, the positive rate of gBRCA1/2 mutations in patients with pancreatic cancer was comparable to that reported previously (6–9). The positive rate of a gBRCA1/2 mutation in patients with a family history of pancreatic, breast, ovarian or prostate cancer is significantly higher than in patients with no such family history. However, accumulated evidence has shown that gBRCA1/2 was detected even in cases with no family history of such cancers (6), Therefore, a risk exists that gBRCA1/2 mutation-positive patients may be omitted if the analysis is limited to only cases with a family history of the aforementioned cancers.

In Japan, mFFX is the only regimen recommended by Japanese Clinical Practice Guidelines as platinum-based chemotherapy for patients with pancreatic cancer (18). Therefore, BRCA analysis is often performed in young adult patients who are eligible for mFFX in clinical practice. In this cohort, the gBRCA1/2 mutation–positive group tended to develop disease at a younger age than the negative group. However, an older patient existed in the positive group, and a statistical significant difference was not observed between the two groups (Table 2, Supplementary Fig. 4). Of note, Momozawa et al. reported that the cumulative risk of pancreatic carcinogenesis in carriers of a gBRCA1/2 mutation increased with age (8). In addition, the NCCN guidelines (19) recommends a FOLFOX regimen for even elderly patients. To improve the OS of patients with pancreatic cancer harbouring a gBRCA1/2 mutation, investigations should be made into a FOLFOX regimen as a standard chemotherapy in Japan. In this study, FOLFOX resulted in tumour shrinkage in an elderly patient with a gBRCA2 mutation; this patient could then be switched to olaparib. Based on our observations, BRCA analysis should be performed in all patients with pancreatic cancer in which chemotherapy is indicated, regardless of age or family history.

As previously reported (4,14), the TRAEs of olaparib were less severe than that those of the platinum-based chemotherapy. Olaparib is therefore suitable for patients who have difficulty with the long-term use of platinum-based chemotherapy. Although it should be noted that no definitive data exists showing that olaparib maintenance therapy contributes to OS, from the perspective of the TRAEs profile of each therapy, patients who are gBRCA1/2 positive should be switched to olaparib maintenance therapy.

The ORR of the platinum-based chemotherapy was superior in gBRCA1/2 mutation-positive patients despite its use as second-line therapy. In contrast, the efficacy of olaparib in our study tended to be worse than that of a previous report (Fig. 2). However, it is necessary to take into account that olaparib was used as second-line therapy after a platinum regimen in the previous study. Two gBRCA1/2 mutation-positive patients in this study were administered platinum-based chemotherapy for ˃4 months prior to switching to olaparib. It has been reported that the long-term use of olaparib and platinum agents causes reversion mutation—a reversion of the function of PARP through a similar mechanism—resulting in resistance to olaparib and platinum agents (20,21). This may have affected the response to olaparib by patients under long-term use of the platinum-based chemotherapy.

Assessing the genetic predisposition of a patient’s family members based on the results of the patient’s BRCA analysis is useful in that it can lead to the surveillance of carriers. From a patient’s perspective, they may feel a sense of guilt towards their blood relatives due to the latter’s possible inheritance of the cancer.

Considering the severe mental and physical strains patients with advanced-stage pancreatic cancer are under, BRCA analysis is expected to be performed at an early stage not only to determine a patient’s treatment strategy, but also for its use in any genetic counselling.

The limitations of this study are as follows: First, the small number of patients recruited from a limited number of hospitals may have introduced a selection bias. Although it is more accurate to obtain informed consent at the time of a diagnosis, in this study the timing of the informed consent obtained varied, including at the time of diagnosis and of testing. This may have also introduced a selection bias. Second, the choice of dosing schedules and the frequency of chemotherapy were left to the attending physician. This may have affected results related to chemotherapy. Third, the observation period was insufficient to assess the OS of gBRCA1/2-positive patients. Despite such limitations, consideration should be given to applying our findings in daily clinical practice in the treatment of patients with unresectable pancreatic cancer.

In conclusion, our prospective multicenter study revealed that platinum-based chemotherapy was effective for all patients harbouring a gBRCA1/2 mutation with advanced and/or metastatic pancreatic cancer, even if its regimen was administered as second-line therapy. Subsequently, all patients were able to switch to active olaparib maintenance therapy. Accordingly, gBRCA1/2 analysis has benefits in the decision-making on therapeutic strategies for patients with pancreatic cancer, regardless of a family history of cancer or age.

Conflict of interest statement

The authors declare that they have no conflicts of interest to report.

Funding

The authors declare that they did not receive any specific funding for this study.