Predictive factors for the diagnosis of autoimmune pancreatitis using endoscopic ultrasound-guided tissue acquisition: a retrospective study in Japan
Article information
Abstract
Background/Aims
The factors affecting the detection rate of lymphoplasmacytic sclerosing pancreatitis (LPSP) using endoscopic ultrasound-guided tissue acquisition (EUS-TA) in patients with type 1 autoimmune pancreatitis (AIP) have not been thoroughly studied. Therefore, we conducted a retrospective study to identify the predictive factors for histologically detecting level 1 or 2 LPSP using EUS-TA.
Methods
Fifty patients with AIP were included in this study, and the primary outcome measures were the predictive factors for histologically detecting level 1 or 2 LPSP using EUS-TA.
Results
Multivariate analysis identified the use of fine needle biopsy (FNB) needles as a significant predictive factor for LPSP detection (odds ratio, 15.1; 95% confidence interval, 1.62–141; p=0.02). The rate of good-quality specimens (specimen adequacy score ≥4) was significantly higher for the FNB needle group than for the fine needle aspiration (FNA) needle group (97% vs. 56%; p<0.01), and the FNB needle group required significantly fewer needle passes than the FNA needle group (median, 2 vs. 3; p<0.01).
Conclusions
The use of FNB needles was the most important factor for the histological confirmation of LPSP using EUS-TA in patients with type 1 AIP.
INTRODUCTION
Autoimmune pancreatitis (AIP) is an atypical type of chronic pancreatitis that is characterized by autoimmune-related infiltration of abundant lymphocytes.1,2 AIP is histopathologically classified as type 1 or 2. Type 1 AIP is an immunoglobulin (Ig) G4-related disease3 that is characterized by histologic features described as lymphoplasmacytic sclerosing pancreatitis (LPSP).4 Type 2 AIP, which is histologically identified as idiopathic duct-centric chronic pancreatitis, is characterized by granulocyte epithelial lesions5,6 and is rare in East Asia compared to Western countries.
The International Consensus Diagnostic Criteria (ICDC) defines AIP based on the characteristic shapes of the pancreatic parenchyma and duct on imaging, serology, other organ involvement (OOI), pancreatic histology, and response to steroid trial.5 Among these factors, the histology is the only factor that can be used by itself for a definitive diagnosis of AIP. In particular, histological evaluation of the involvement of pancreatic malignancies is mandatory for focal AIP lesions. Although obtaining sufficient specimens for the histological evaluation of AIP without undergoing surgery is generally difficult, reports have recently demonstrated the sensitivity of endoscopic ultrasound-guided fine needle aspiration (EUS-FNA) for the histological diagnosis of AIP, due to improved EUS-FNA specimen processing7-9 and newly-designed EUS-FNA needles.10-12 Furthermore, the development of EUS-guided fine needle biopsy (EUS-FNB) needles has enabled a breakthrough in the histological diagnosis of AIP.10-14 Therefore, tissue acquisition for histological evaluation has become increasingly important when diagnosing AIP.
However, the sensitivity of EUS-tissue acquisition (EUS-TA) for the histological detection of LPSP may be affected by several factors, and predictive factors for the histological detection of LPSP using EUS-TA in patients with type 1 AIP are necessary. Thus, we conducted a retrospective study to clarify the predictive factors for histologically detecting levels 1 and 2 LPSP using EUS-TA.
METHODS
Study population and data collection
Among the 100 patients who were retrospectively diagnosed with AIP based on the ICDC criteria using the medical records of our medical center between January 2002 and December 2023, 50 who underwent EUS-TA for AIP lesions were included in this study. Laboratory data, imaging findings, and information on the EUS-TA procedures were retrospectively collected. We obtained the EUS-TA-based histological diagnoses from the registered pathological database. Since 2011, all patients with AIP have undergone EUS-TA procedures in our hospital. Therefore, all diagnoses were based on the 2011 ICDC.
Outcome measurements
In this study, the primary outcome measures were the predictive factors for the histological detection of levels 1 and 2 LPSP using EUS-TA.
EUS-TA procedure
The endoscopic and ultrasonography equipment used were GF-UCT260, EU-ME1, and EU-ME2 (Olympus Co.). Although we previously used 19G or 22G conventional EUS-FNA needles to perform EUS-TA for lesions suspected to be AIP, we have primarily used 22G EUS-FNB needles for the histological diagnosis of AIP since 2017. Each needle pass involved 5 to 20 strokes with suction using a 20 mL syringe. Although at least two needle passes were usually performed during EUS-TA for suspected AIP lesions, additional passes were performed when the obtained specimen volumes were macroscopically inadequate. All EUS-TA procedures were performed by an expert endosonographer or trainee under the supervision of experts.
When conventional EUS-FNA needles were used, most specimens inside the needles were pushed out from the inside of the needles using a small volume of saline or a needle stylet. Specimens were pushed into a centrifuge tube containing 5 mL of a 10% formalin solution and were processed using the cell block method.15 Since 2017, when EUS-FNB needles were used, macroscopic on-site evaluation (MOSE)16 was used to initially confirm whether an adequate volume of whitish tissues was obtainable; the remaining specimens that were not used for MOSE were separately processed using the cell block method. Rapid on-site evaluation was not performed.
Histocytological evaluations
All histocytological diagnoses of the specimens obtained using EUS-TA were made based on the consensus of at least two gastrointestinal pathology experts. All EUS-TA specimens, including cell block sections, were first stained with hematoxylin and eosin (H&E) and periodic acid-Schiff and Alcian blue. If clinical and histological findings from H&E staining suggested type 1 AIP, additional immunostaining was performed using antibodies against IgG4 (HP6023; Millipore) and leukocyte common antigen (2B11+PD7/26; Nichirei). Elastica-Masson staining was performed to assess fibrosis and obliterative phlebitis (OP). When pancreatic malignancies, including pancreatic ductal adenocarcinomas, were suspected based on histological findings from H&E staining, antibodies against Ki67 (MIB-1; Immunotech), P53 (DO-7; DAKO), and MUC1 (Ma695; Novocastra) were used for further evaluation.
For EUS-TA specimens, LPSP was diagnosed using the ICDC.5 Level 1 LPSP was defined as the presence of 3 or more of the following four histological features: (1) lymphoplasmacytic infiltration (LI), (2) IgG4-positive plasma cell infiltration (IgG4-positive PI) with >10 counts/high-power field (abundant IgG4-positive PI), (3) presence of storiform fibrosis (SF), and (4) OP.3 Level 2 LPSP was diagnosed when two of the aforementioned histological features were present.
Specimen adequacy was categorized using the scoring system described by Gerke et al.17: 0, insufficient material for interpretation; 1, sufficient material for limited cytologic interpretation; 2, sufficient material for adequate cytologic interpretation; 3, sufficient material for limited histologic interpretation; 4, sufficient material for proper histological interpretation of low quality (total material within a ×10 power field in length); and 5, sufficient material for adequate histological interpretation of high quality (total material ˃×10 power fields in length). In this study, EUS-TA specimens with scores ≥4 were defined as good-quality specimens.
Statistical analyses
Categorical variables were compared between the groups using Fisher exact test, and continuous variables were compared using the Mann-Whitney U-test. Multivariate analysis was performed using a multiple logistic regression model. Factors with a p-values <0.05 in the univariate analyses were further evaluated using a multiple logistic regression model; odds ratios (ORs) and 95% confidence intervals (CIs) were calculated. Statistical significance was set at p<0.05. All statistical analyses were performed using IBM SPSS Statistics software ver. 24.0 (IBM Corp.).
Ethical statements
This study was approved by the Sendai City Medical Center Institutional Review Board (registration number: 2023-0035).
RESULTS
Patient characteristics
Table 1 shows the baseline characteristics of the patients with AIP who underwent EUS-TA. The 50 patients with AIP comprised 38 males and 12 females. When patients underwent EUS-TA, their mean age was 71.0±8.8 years. Serum IgG4 levels of ≥135 mg/dL were detected for 39 patients (78.0%), and seven patients (14.0%) had OOI. Imaging studies, including contrast-enhanced computed tomography, performed immediately before EUS-TA detected diffuse pancreatic enlargement in 48.0% (24/50) of the patients. Based on the ICDC, 48 patients (96.0%) were diagnosed with definitive/probable type 1 AIP, and the remaining two were diagnosed with AIP not otherwise specified. After EUS-TA for AIP lesions, 33 patients (66.0%) received oral prednisolone as standard therapy for AIP, and 16 (32.0%) underwent surveillance without standard steroid therapy. The remaining patient underwent surgery because of EUS-TA-based diagnosis of suspected adenocarcinoma.
Baseline characteristics of AIP patients who received EUS-TA
EUS-TA procedures for AIP lesions and histological results
Details of the EUS-TA used in this study are shown in Table 2. EUS-FNB needles were used in 64.0% of AIP patients with AIP (32/50). A 22G needle was primarily used for EUS-TA (80%). No adverse events related to the EUS-TA procedure were observed.
Details of the EUS-TA procedures
The histological results of the EUS-TA specimens obtained from AIP lesions are shown in Table 3. Level 1 LPSP (Fig. 1) was histologically confirmed in 33 (66.0%), and level 1 or 2 LPSP was confirmed in 41 (82.0%) patients; fibrotic benign tissues that did not meet level 1 or 2 LPSP criteria were obtained via EUS-TA from eight patients (16.0%). Fibrotic tissues with suspected malignancy were obtained from the remaining patient (2%).
Results from EUS-TA for patients with AIP
Histological findings of the specimens obtained using endoscopic ultrasound-guided fine needle biopsy. Hematoxylin and eosin staining of the specimens shows abundant lymphoplasmacytic infiltration and storiform fibrosis (A, ×100). Immunoglobulin (Ig) G4 immunostaining shows abundant IgG4-positive plasma cells (B, ×400). Elastica-Masson staining shows obliterative phlebitis (C, ×200).
Predictive factors for the diagnosis of level 1 or 2 LPSP using EUS-TA specimens
Table 4 shows the univariate and multivariate analytical results of the predictive factors for diagnosing level 1 or 2 LPSP using EUS-TA specimens. We identified 10 candidate predictive factors for the diagnosis of LPSP using EUS-TA. Of the 10 candidate factors, two (use of FNB needles and MOSE) had p-values <0.05, as determined by univariate analysis. An exploratory multivariate analysis was performed, and the use of FNB needles was identified as a significant predictive factor for LPSP detection (OR, 15.1; 95% CI, 1.62–141.0; p=0.02).
Predictive factors for the diagnosis of level 1 or 2 LPSP
Comparison of the diagnostic abilities of FNB and FNA needles
Since the use of FNB needles was identified as a critical factor in obtaining a histological diagnosis of LPSP using EUS-TA, we investigated the differences in EUS-TA procedures and histological findings of EUS-TA specimens between the FNB (n=32) and FNA (n=18) needle groups. Despite the significantly lower number of needle passes in the FNB needle group than in the FNA needle group (median, 2 vs. 3; p<0.01), the histological results of the EUS-TA specimens were significantly better in the FNB needle group than in the FNA needle group (Table 5). The rate of good-quality specimens (specimen adequacy score ≥4) was significantly higher for the FNB needle group than for the FNA needle group (97% vs. 56%, p<0.01). Histological confirmation of LPSP (level 1 or 2) was significantly more frequent in the FNB needle group than in the FNA needle group (97% vs. 56%, p<0.01). The detection rates of LI and abundant IgG4-positive PI were significantly higher in the FNB needle group than in the FNA needle group (97% vs. 56%, p<0.01 and 94% vs. 61%, p<0.01, respectively).
Comparison of procedural and histological fndings for FNB and FNA needles
DISCUSSION
Although several factors affecting the diagnostic accuracy of EUS-TA have been identified, the key factors for the histological confirmation of LPSP using EUS-TA required clarification. This study was conducted in real-world clinical practice, and the results demonstrated that the use of EUS-FNB needles was the most important factor for the histological confirmation of level 1 or 2 LPSP using EUS-TA in patients with clinically suspected AIP.
Prospective clinical studies have clarified the excellent ability of EUS-FNB to diagnose LPSP in patients with suspected AIP.10-12 Pooled data from two meta-analyses indicated that the diagnostic yield of FNB needles was better than that of FNA needles for the diagnosis of AIP, due to the ability to obtain an adequate amount of tissues while maintaining the histological structures.13,14 In this study, however, we examined several candidate factors affecting the confirmation of LPSP using EUS-TA, including those related to the EUS-TA procedures and devices, and the use of FNB needles had the most significant impact on the diagnosis of LPSP. In addition, our study showed that the rate of good-quality specimens (specimen adequacy score ≥4) was significantly higher for the FNB needle group than for the conventional FNA needle group, which may explain the significant impact of FNB needles on the detection rate of LPSP. Our study and a previous study14 have shown that fewer needle passes are required for the histological evaluation when FNB needles are used. This implies that the diagnostic ability of FNB needles does not depend on the number of needle passes. However, the ability of conventional EUS-FNA to detect malignancy has been reported to gradually improve with up to six needle passes.18 In our study, the number of needle passes (≥4) was not a significant factor in the diagnosis of LPSP.
In this study, we investigated the detection rates of four histological factors associated with the diagnosis of LPSP using EUS-TA. The detection rates of LI, abundant IgG4-positive PI, and SF were relatively good, ranging from approximately 70% to 80%, whereas the detection rate of OP was low (<25%). In addition, the respective detection rates of LI and abundant IgG4-positive PI using FNB needles were excellent (>90%) and significantly higher than those obtained using conventional FNA needles, which may have contributed substantially to the high LPSP detection rate using FNB needles. In contrast, the detection rates of SF and OP were not significantly different between the FNB and conventional FNA groups. This can be explained by the heterogeneous distribution of SF and OP within AIP lesions; the detection of SF and OP may depend on chance. What is worse, those two factors are sometimes equivocal, even in large amounts of tissue.19 Although a larger needle size seems to improve the detection rate of LPSP using EUS-TA, a previous study14 and our study found no evidence of the superiority of a larger needle compared with a conventional needle size (22G or smaller) for the detection of LPSP. To improve the ability of EUS-TA to diagnose AIP lesions, factors contributing to the improvement in the detection rates of SF and OP should be studied in the future.
The results of this study showed that the detection rates of LPSP using EUS-TA were the same for diffuse and focal morphologies (88% vs. 77%, p=0.47). Recently, we reported that EUS-TA plays an important role in diagnosing focal AIP because EUS-TA is required to diagnose AIP in approximately 80% of patients with focal AIP.20 Since most patients with diffuse AIP can be diagnosed as having AIP using serological and imaging findings, the use of FNB needles is meaningful, especially for patients with focal AIP.21 Therefore, when pancreatic masses are suspected of AIP based on imaging findings, including homogenous delayed enhancement,20 FNB needles should be used.
This study had some limitations. First, it was retrospectively conducted at a single institution and involved a small sample size. Therefore, large-scale validation studies are warranted. Second, the needles were selected at the discretion of the endoscopist, leading to selection bias. Third, EUS-TA procedures and sample processing methods not commonly performed in this study, such as the slow-pull technique, were not evaluated. Therefore, there may be predictive factors for the diagnosis of LPSP using EUS-TA that were not investigated in this study. In addition, operator-dependent factors could not be evaluated because two or more endosonographers, including trainees and experts, often performed EUS-TA for each patient. Fourth, because most patients who underwent MOSE were categorized into the FNB needle group, the high detection rate of LPSP using EUS-TA may have partially relied on the additional effect of MOSE. Although the multivariate analysis did not identify MOSE as a significant factor affecting the detection rate of LPSP using EUS-TA, the possibility that MOSE is a dependent factor affecting the accuracy of EUS-TA for patients with AIP cannot be ruled out because performing MOSE involves some selection bias. Fifth, no patients were diagnosed with type 2 AIP in this study, and whether EUS-TA is useful in patients with type 2 AIP remains to be determined.
In conclusion, although several clinical and procedural factors affect the accuracy of EUS-TA in detecting the histology and malignancy in several pancreaticobiliary diseases, the use of FNB needles may be the most important factor in the histological confirmation of LPSP using EUS-TA.
Notes
Conflicts of Interest
The authors have no potential conflicts of interest.
Funding
None.
Author Contributions
Conceptualization: KY, SK; Date curation: KY, SK; Formal analysis: KY, SK, YN, KI; Investigation: KY, SK, YK, TO, HK, TS, KM, FK, HO, YM, KH, HS; Methodology: KY, SK; Project administration: KY, SK; Supervision: SK, KI; Writing–original draft: KY, SK; Writing–review & editing: all authors.