Diffusion-weighted Imaging: New Paradigm in Diagnosis of Early Acute Pancreatitis.

Acute pancreatitis is a noninfective inflammation of the pancreas characterized by the release of inflammatory mediators and resultant acinar cell death. The disease is characterized by a spectrum of clinical presentation from mild interstitial to severe necrotic forms. The morbidity and mortality ranges from 10% to 40% in varied populations depending on the disease severity and comorbidities of the patient.[ 1 2 3 4 ] The trifecta of clinical profile of upper abdominal pain, laboratory parameters of increased serum enzymes, and radiological evidence form the cornerstone of diagnosis. However, abdominal pain may be vague or absent in up to 10% of patients, and laboratory parameters may not be helpful in around 20% of cases, especially in early and mild disease. In this scenario, the onus of diagnosis lies with radiology. Contrast-enhanced computed tomography (CECT) abdomen is the currently prevalent radiological investigation for diagnosis of acute pancreatitis and its complications. Although highly sensitive and specific, it is limited by the lesser efficacy in the diagnosis of early and mild interstitial cases. Hence, in mild acute pancreatitis, all three diagnostic criteria may fail to impress upon the diagnosis.[ 5 6 7 ] Therefore, a lacuna for a diagnostic modality rises for early pancreatitis, which can pave the way for timely management and prevention of disease progression to more severe form. In such settings, magnetic resonance imaging (MRI) with diffusion-weighted imaging (DWI) has shown promise in recent years due to its rapid implementation, the absence of radiation or iodinated contrast, and high-contrast resolution. DWI analysis of random molecular motion of water across cells, also known as Brownian motion or diffusion. A lack of such movement in the scenario of cell death or inflammatory substance deposition results in a higher intensity are on high b values on DWI giving an appearance of restricted diffusion in comparison to normal tissue. Further objective analysis may be done by apparent diffusion coefficient (ADC) which is a quantitative measurement of Brownian motion within the tissues. The cellular microenvironment is reflected by ADC quantification where inflamed or nonviable cells show lower values in comparison to healthy cells.[ 8 9 10 ] In the present study, it was aimed to study the diagnostic accuracy of DWI in the diagnosis of early acute interstitial pancreatitis in 50 cases of clinically diagnosed pancreatitis in comparison with 50 controls.

The study group comprised 50 patients with 29 (58%) females and 21 (42%) males. The mean age of both genders suggested that males (mean age = 33.85 years) presented with features of pancreatitis about a decade earlier than females (mean age = 45.41 years). Almost three-fourth of female patients presented with biliary causes of pancreatitis, whereas only 42.8% of males presented with known biliary disease. Other/unknown causes prevailed as etiological factors in the male group. Among the cases, 48 out of 50 patients showed restricted diffusion on qualitative analysis, whereas none of the controls showed restricted diffusion. On quantitative evaluation of mean ADC values, the mean values in the cases (0.882 × 10 −3 mm 2 /s ± 0.102) were significantly lower than the control group (1.178 × 10 −3 mm 2 /s ± 0.232) with P < 0.001. However, the ADC values showed no significant difference across genders or etiological causes. Analysis with the ROC curve revealed an area under curve of 0.87 suggesting an excellent efficacy of DWI in diagnosing acute pancreatitis. The ROC curve deciphered a cutoff value of 0.937 × 10 −3 mm 2 /s for diagnosis of acute pancreatitis with quantitative evaluation of ADC values. This cutoff value has a sensitivity of 84.00% and a specificity of 86.00% in the present study population. The PPV of ADC values is 85.71% with an overall diagnostic accuracy of 85.00% in predicting acute interstitial pancreatitis. Table 1 depicts the demographic and etiological characteristics of the study group. Table 2 reveals the relationship between the mean ADC values of both the groups. Figure 1 shows the ROC curve drawn for study parameters. Table 3 shows the statistical analysis drawn by the ROC curve. Table 4 depicts the relationship between gender and etiology with ADC values. A few representative cases are shown in Figures 2 - 4 . Receiver operating characteristic curve using apparent diffusion coefficient value for differentiating between cases and controls. ROC: Receiver operating characteristic Apparent diffusion coefficient (ADC) mapping in a 45-year-old female with clinical features of acute pancreatitis shows mean ADC values of 996 × 10 −6 mm 2 /s (head), 922 × 10 −6 mm 2 /s (body), and 938 × 10 −6 mm 2 /s (tail), with an overall mean of 952 × 10 −6 mm 2 /s Contrast-enhanced computed tomography abdomen axial images (a) of a 10-year-old boy with diffusely bulky pancreatic body and tail corresponding with diffusion restriction on magnetic resonance imaging (b) in the same region Contrast-enhanced computed tomography images (a and b) show bulky and irregular pancreas with peripancreatic fluid collection and inflammation. Corresponding axial apparent diffusion coefficient (ADC) maps (c and d) show a mean ADC of 936 × 10 −6 mm 2 /s in the inflamed pancreatic head and 974 × 10 −6 mm 2 /s in the inflamed pancreatic tail Demographic and etiological characters of the study group Relationship between mean apparent diffusion coefficient values of both groups ADC=Apparent diffusion coefficient Statistical analysis derived from the receiver operating characteristic curve PPV=Positive predictive value, NPV=Negative predictive value, AUC=Area under curve, ROC=Receiver operating characteristic, CI=Confidence interval Relationship between sex/cause and mean apparent diffusion coefficient value in the patient group ADC=Apparent diffusion coefficient, SD=Standard deviation, BC=Biliary causes

MRI evaluation with qualitative and quantitative DWI can thus be a succinct modality for early diagnosis of pancreatitis with excellent diagnostic accuracy and superlative advantage of lack of contrast and radiation in comparison to a CT scan. Along with pancreatitis, DWI has the propensity of differentiating other pancreatic parenchymal diseases as well as simultaneous evaluation of biliary calculus disease. Hence, a paradigm shift to MRI from CECT for the evaluation of pancreatitis is propounded. The authors certify that they have obtained all appropriate patient consent forms. In the form, the patient(s) has/have given his/her/their consent for his/her/their images and other clinical information to be reported in the journal. The patients understand that their names and initials will not be published and due efforts will be made to conceal their identity, but anonymity cannot be guaranteed. Nil. There are no conflicts of interest.

Acute pancreatitis characterized by cell death and inflammatory substances deposition leads to restricted diffusion with lower ADC values in the inflamed region on DWI. These features make DWI a potentially useful tool in early diagnosis. However, with CECT abdomen being the standard investigation for pancreatitis and its associated features, a paradigm shift to noncontrast MRI with DWI is yet to take shape. The higher costs, relatively limited availability, and longer duration of scanning had put MRI on the back foot in the past. Albeit, recent advances in MRI have allowed ultrafast imaging and reduced imaging artifacts caused by motion. Furthermore, the ability of early diagnosis and no threat in patients with compromised renal function have brought MRI into the clinical frame of management in recent years.[ 11 12 13 ] Shinya et al . and Wiggerman et al . were among the first researchers to establish that ADC values can differentiate between normal and diseased pancreas.[ 12 14 ] Over the years, the utility of DWI has been studied in assessing its role in differentiating pancreatic lesions of varied etiopathogenesis including pancreatic inflammation and pancreatic neoplasms. The present study group showed female preponderance with 58% of female participants. Thirty out of 50 patients of the study population (60%) presented with biliary causes of pancreatitis. This was comparable to the etiological profile of the study group of Salik et al . who reported a female presence of 73% in their study group of 63 patients with acute pancreatitis. They also reported a similar prevalence of biliary disease in their study group (62%).[ 15 ] Li et al . while evaluating MRI and DWI findings of 66 patients with acute pancreatitis reported that 50% of the study population had known biliary causes, whereas 41% presented with unknown causes.[ 16 ] In the present study, it was observed that 48 out of 50 patients of acute pancreatitis in early stages showed diffusion restriction on qualitative assessment, which was significantly different than the control group. Quantitative ADC evaluation revealed a significant difference in the mean values of cases and controls. Abuzeid et al., while studying 16 patients of acute pancreatitis with a similar methodology also deciphered that the mean ADC values in cases of pancreatitis were 1.15 × 10 −3 mm 2 /s while that of controls was 1.60 × 10 −3 mm 2 /s.[ 11 ] Li et al . also reported a significant difference in mean ADC value of 1.68 ± 0.45 × 10 −3 mm 2 /s in 66 patients of acute pancreatitis in comparison to that of 2.09 ± 0.55 × 10 −3 mm 2 /s in the control group.[ 16 ] Thomas et al ., while studying the diffusion MRI features of 27 patients of acute pancreatitis and 38 controls, reported a mean ADC of 1.32 × 10 −3 mm 2 /s in patients and 1.77 × 10 −3 mm 2 /s in controls.[ 17 ] Nam et al . assessed qualitative diffusion on similar criteria of comparison with that of the spleen as in the present study. They compared computed tomography (CT) and MRI findings of 85 patients with acute pancreatitis and presented a mean ADC value of 1.09 × 10 −3 mm 2 /s in patients in comparison to 1.46 × 10 −3 mm 2 /s in controls.[ 18 ] Thus, a uniformity in recent literature has been observed in depicting mean ADC values of normal and acutely inflamed pancreas. After depicting a significant difference in quantitative evaluation of mean ADC values, a cutoff value of 0.937 × 10 −3 mm 2 /s was derived for diagnosis of acute interstitial pancreatitis on DWI as per the findings of the current study group with a sensitivity of 84% and specificity of 86%. The nearest comparative ADC threshold was provided by Kamisawa et al . as 1.075 × 10 −3 mm 2 /s in the diagnosis of autoimmune pancreatitis on DWI.[ 19 ] In comparison to this, Tonolini and Di Pietro et al . presented a cutoff ADC value of 1.1 × 10 −3 mm 2 /s and Thomas et al . proposed a threshold ADC value of 1.6 × 10 −3 mm 2 /s with a sensitivity of 93% and specificity of 87% in diagnoses.[ 5 17 ] A diagnostic accuracy of more than 90% in the present study is well in line with various other studies, thereby paving the path of DWI as a major diagnostic player in acute pancreatitis. While the present study focused only on the diagnostic evaluation, attempts have been made to correlate ADC values with disease severity on the scale of clinical scoring systems. Hocaoglu et al . studied the ADC values in pancreatitis patients with mild pancreatitis with a Ranson score of 1–3. They emphasized a significant reduction in mean ADC in patients of early pancreatitis in comparison to normal pancreas, thereby suggesting that DWI improves diagnosis at early stages of the disease.[ 8 ] While comparing the ADC values with clinical Balthazar scoring, Yencilek et al . observed a significant correlation of ADC with disease severity. The mean ADC in their study group ranged from 1.45 × 10 −3 mm 2 /s ± 0.24 to 0.89 × 10 −3 mm 2 /s ± 0.18 from Grade A to E of Balthazar staging.[ 6 ] In a similar evaluation, Li et al . reported a decreasing trend in mean ADC values from mild-to-severe pancreatitis.[ 16 ] İlhan et al . deduced a comparable efficacy of DWI with CECT abdomen and clinical grading systems: Balthazar and Revised Atlanta scoring.[ 13 ] The scope of using DWI in pancreatitis has also been researched for other varieties of pancreatic inflammation including mass-forming pancreatitis, autoimmune pancreatitis, and chronic pancreatitis. Since all these varieties are often imaging mimickers of pancreatic malignancy, DWI is poised as the distinguisher between various nonmalignant and malignant lesions. Kamisawa et al . studied ADC values in 13 patients of autoimmune pancreatitis and 40 patients of pancreatic neoplasms, obtaining significant differences in ADC values of inflamed and neoplastic pancreas. A cutoff threshold of 1.075 × 10 −3 mm 2 /s was proposed by them to differentiate autoimmune pancreatitis from neoplasms.[ 19 ] Akisik et al . tested the efficacy of DWI in differentiating chronic pancreatitis from normal parenchyma postsecretin stimulation revealing lower baseline ADC which corresponded with clinical severity.[ 20 ] El-Shinnawy et al ., on the other hand, suggested that pancreatic neoplasms had lower ADC values in comparison to mass-forming pancreatitis, thereby providing a diagnostic delineation.[ 21 ] Recently, Schima et al . also deciphered on similar lines that on 3.0 T MRI, the mean ADC value of pancreatic ductal carcinomas was significantly lower than that of mass-forming pancreatitis (1.17 × 10 −3 mm 2 /s ± 0.23 vs. 1.47 × 10 −3 mm 2 /s ± 0.18, P < 0.01).[ 22 ] Furthermore, for assessment of pancreatic malignancies, a recent meta-analysis suggested that DWI evaluation is equally efficacious as positron emission tomography scan.[ 23 ] A moderate sample size and retrospective study design are the major limitations of the present study. A correlation of laboratory values and clinical staging with pancreatic ADC values was thus precluded and gradation of ADC in comparison to clinical severity could not be obtained. Furthermore, a larger sample size, a multicenter trial, or a meta-analysis that can provide a uniform ADC value threshold across platforms will enhance the diagnostic utility of the modality overcoming the varied numerical cutoffs across regions.

The present retrospective observational study was conducted at the department of radio-diagnosis of a tertiary teaching hospital in North India. The study is based on the imaging data of patients suffering from early acute pancreatitis who presented for MRI evaluation for a period of one year from July 2022 to June 2023. Institutional ethical committee approval was obtained through letter no SGRR/IEC/04/23. All the procedures followed were in accordance with the Declaration of Helsinki 1975, as revised in 2013 or similar standards. Patients of all age groups underwent MRI imaging of the pancreas with clinical or laboratory diagnosis of early pancreatitis. Patients of all age groups underwent MRI imaging of the pancreas with clinical or laboratory diagnosis of early pancreatitis. Acute necrotic pancreatitis or acute on chronic pancreatitis Pancreatitis in association with pancreatic malignancy Diffusion-weighted image quality precluding accurate measurement of ADC values. Acute necrotic pancreatitis or acute on chronic pancreatitis Pancreatitis in association with pancreatic malignancy Diffusion-weighted image quality precluding accurate measurement of ADC values. Age- and gender-matched patients undergoing MRI abdomen/magnetic resonance cholangiopancreatography with radiologically normal pancreas and no clinical suspicion of pancreatitis. MRI data acquisition was conducted utilizing a 1.5 Tesla scanner (Magnetom, Siemens, Erlangen, Germany) equipped with a standard body coil for both conventional imaging and DWI. The imaging protocol encompassed axial and coronal T2-weighted turbo spin-echo images, axial and sagittal short-tau inversion recovery images, and axial T1-weighted turbo spin-echo images. In addition, axial DWI was performed employing a multislice single-shot echo-planar imaging sequence, incorporating two b values (0, 800 s/mm 2 ). Key DWI acquisition parameters included a repetition time of 1400 ms, echo time of 70 ms, section thickness of 6 mm, and field of view of 380 cm. Furthermore, to complement DWI findings, ADC maps were automatically generated, utilizing a monoexponential decay model incorporating the b values. A radiology consultant with 12 years of experience in reporting MRI abdomen evaluated all images for DWI and ADC analysis. The reporting radiologist was blinded to laboratory results. Subjective evaluation of diffusion restriction was done by taking splenic parenchyma as a reference. Pancreatic tissue brighter than the spleen on DWI was considered restricted on DWI. This was further confirmed by quantitative comparative values of pancreas and spleen with normal pancreatic parenchyma having higher ADC values than spleen, whereas in patients of pancreatitis, the ADC values were similar or less than that of the spleen. ADC values were calculated in both cases and controls. In controls, ADC values were designated in pancreatic segments of head, body, and tail within a region of interest (ROI) of 50 mm 2 ± 10 mm 2 in each segment. Mean was calculated as the final depictive value of normal pancreas. In cases, where there was diffuse restricted diffusion, head, body, and tail segment, ADC values were measured with similar ROI to that of controls. In cases of focal pancreatitis, ROI was placed in the involved pancreatic segment, and three values were obtained from that particular segment. Mean ADC values were calculated in all cases from the obtained three values. Areas with fluid, necrosis, collections, ducts, pseudocysts, cysts, and artifacts were excluded from the study. Categorical variables were assessed as percentages. The measures of central tendency in nominal variables were studied as mean. Tests of association were performed in the form of t -test and ANOVA. Tests of diagnostic accuracy were calculated using the receiver operating characteristic (ROC) curve to determine cutoff ADC values for diagnosis of acute pancreatitis as well as sensitivity, specificity, positive predictive value (PPV), and negative predictive value of DWI for diagnosis of acute pancreatitis. P < 0.05 was taken as significant. All calculations were done using SPSS version 27 (IBM SPSS Statistics for Windows, Version 27.0, Armonk, NY: IBM Corp).