Ultrasound point shear wave elastography of the pancreas: comparison of patients with type 1 diabetes and healthy volunteers – results from a pilot study

Type 1 diabetes (T1D) is an autoimmune disease in which insulin-producing beta cells in the pancreas are destroyed. Comparative international figures show an annual increase of 3–4% in the global incidence rates of T1D [1]. Since there is a strong genetic contribution to T1D, further screening modalities could be of interest for patients at genetic risk [2]. Shear wave elastography presents a method for determining the elasticity of the target tissue, using ultrasound to determine the velocity of shear waves [3]. The procedure was originally developed as a non-invasive investigation for the early diagnosis and treatment of liver fibrosis [4]. D’Onofrio et al. and Kawada et al. described ductal adenocarcinoma as stiffened tissue with an increased shear wave velocity (Vs) due to the fibrotic process [3, 4]. Studies have also shown an increase in the Vs in patients with chronic pancreatitis [5]. Recent studies on type 2 diabetes (T2D) have demonstrated that elastographic examination of the pancreas shows an increased Vs in affected patients compared with healthy subjects [6]. Concerning type 1 diabetes, a recent study on children and adolescents did not show any increased values in those with T1D. However, that study looked only at the body of the pancreas [7].

It is assumed that pathological mechanisms associated with fibrotic change cause the inflamed tissue to increase stiffness [3‐5]. As inflammation in the islets of Langerhans (insulitis) and fibrosis also occur in T1D [8‐10], questions arise as to whether the inflammatory process and the associated alteration in tissue consistency can be measured by shear wave elastography and whether changes can be detected even in the prediabetic phase, when an insulitic process is already taking place. As the results of the various elastographic studies show a certain lack of agreement regarding both examination precision and current clinical relevance [3, 5], this procedure has still to be established for the diagnostic investigation of pancreatic disease.

Although a growing understanding of the pancreatic pathology is allowing the development of novel immune intervention strategies to alter the course of insulitis, additional non-invasive diagnostic tools are still desirable for the early diagnosis of T1D and monitoring the course of disease [2].

The aims of this pilot study were therefore to establish shear wave elastography in the three anatomical regions of the pancreas (head, body, and tail) and evaluate possible alterations in stiffness due to the pathology of T1D by comparing elastographic measurements in patients with T1D and healthy controls.

The shear wave velocity of the tail differed significantly between patients with T1D and healthy volunteers indicating a higher stiffness in T1D (p = 0.0474) (Table 2). In contrast, the Vs in the head and body was not significantly different between the two groups. With respect to the elastographic measurements, it was also of interest to establish the segment of the organ showing the highest and lowest velocity in each group. Comparing the means, we found the highest velocity in the body of the pancreas in both groups (1.2 ± 0.2).

In the control group of healthy volunteers, we found a positive correlation with age for the individual pancreatic segments of the head and body, as well as for the whole pancreas (head: p = 0.0009, R = 0.76308; body: p = 0.0022, R = 0.72589; whole pancreas: p = 0.0012, R = 0.75430) but established no such relationships in the patient group. There was no association between Vs and duration of diabetes in the study group for either of the individual pancreatic segments or the whole organ (p = 0.7534) nor was there any correlation between BMI and Vs in either group (patients: p = 0.7129; controls: p = 0.5402). The position of the pancreas in the abdomen meant that the maximum ultrasound penetration depth of 8 cm was insufficient in one patient and made it impossible to take any measurements in the deeper-lying segments. Large quantities of intestinal air can also make it more difficult or even impossible to measure the pancreas. For this reason, the tail segment could not be measured in two of our patients.

Both ultrasound strain elastography and ultrasound shear wave elastography appear to play an increasing role in the diagnostic investigation of pancreatic disease [13, 14]. Despite the increasing importance of elastographic procedures in the diagnosis of pancreatic disease, we are aware of only a few studies on shear wave elastography of the pancreas in patients with diabetes [15, 16]. In our pilot study, we intended to demonstrate the feasibility of measuring all three anatomical parts of the pancreas and determine measurement limitations with respect to patient characteristics and examination conditions.

Our results (Table 2) show a significant difference between patients and healthy volunteers in the tail of the pancreas (p = 0.0474). The mean value for all segments showed only a trend towards a higher Vs in the patient group.

In a recent study, Saglam et al. did not find any significant differences between children and adolescents with T1D and healthy subjects, but the measurements were performed only on the body of the pancreas. This finding agrees with our results for that region. They reported Vs of 1.09 ± 0.22 m/s for the body of the pancreas in healthy control subjects and 0.99 ± 0.25 m/s for patients with type 1 diabetes – values slightly lower than we found in our patients [7].

Chronic inflammatory processes lead to fibrotic remodelling consistent with non-physiological wound healing not only in the islets of Langerhans (insulitis) but also in the exocrine tissue of the pancreas [17‐19]. Pancreatic fibrosis particularly affects patients with longstanding disease [10, 20]. In our study, the longest duration of disease was about 10 years.

The higher velocities detected may not be due exclusively to fibrosis, as other histological changes may impact the physical properties of the tissue. But fibrotic changes are known to result from type 1 diabetes [17, 21, 22].

One possible explanation for our significant results from the tail could be a greater accumulation of inflammatory CD8+ T cells in this organ segment during insulitis. We can assume that these cells are not only responsible for considerable destruction of the beta cells but are also to be found in greater numbers in the exocrine tissue [19]. In the context of recurrent inflammation of the islets of Langerhans in T1D, Rajput et al. described a higher concentration of beta cells in the tail of the pancreas [10]. This would suggest that patients with T1D have stiffer tissue in the tail of the pancreas, giving higher Vs values in this segment.

Harada et al. used shear wave elastography with ultrasound to examine the body of the pancreas preoperatively in patients about to undergo pancreatic resection. They showed a strong correlation of Vs to the severity of pathological fibrosis in the pancreatic tissue [15].

Stumpf et al. found normal values for the head of the pancreas to be 1.44 ± 0.39 m/s for women and 1.19 ± 0.29 m/s for men. Values for the body were 1.49 ± 0.37 m/s for women and 1.26 ± 0.30 m/s for men, while the corresponding figures for the tail were 1.29 ± 0.36 m/s and 1.05 ± 0.30 m/s, respectively [14]. The values given by Harada et al., with 1.35 m/s for the body of the pancreas, are to a large extent in agreement with the values given by Stumpf et al. and our measurements [15]. Their results for male participants agree extremely well with our results. As the percentage of women in our healthy control group was only 30%, this may explain why the values in the literature for the two sexes combined are slightly higher than our results. Yashima et al. found mean values for head, body and tail of 1.23 m/s, 1.3 m/s and 1.24 m/s in healthy volunteers and of 1.65 m/s, 2.09 m/s and 1.68 m/s in patients with chronic pancreatitis [5]. Xie et al. reported mean values for the pancreatic head and body of 1.18 m/s and 1.21 m/s in healthy controls [23]. The data from Goertz et al. are in accordance with these measurements, with a mean of 1.2 m/s. The average for patients with chronic pancreatitis in that study was 2.21 m/s [24]. Mateen et al. reported results of 1.28 m/s, 1.25 m/s, and 3.28 m/s for healthy controls, patients with chronic pancreatitis, and patients with acute pancreatitis, respectively [25]. Llamoza-Torres et al. obtained figures of 1.27 m/s for healthy volunteers and 1.57 m/s for patients with chronic pancreatitis [26]. Further studies evaluate ARFI as a suitable and promising procedure for the non-invasive diagnostic investigation of chronic pancreatitis and other pancreatic conditions, both benign and malignant, such as cystic lesions and carcinomas [12, 26‐29]. In the literature, the highest Vs was recorded in the body of the pancreas whenever the individual pancreatic segments were measured. Our results confirm these findings, as the Vs of the body of the pancreas was significantly higher in both groups.

In the healthy volunteers, we found a very strong relationship between the Vs and the age of the participant, confirming the observation of Stumpf et al. [14]. Nevertheless, in contrast to their study, we did not find any correlation between Vs and BMI. Nor did we determine a correlation between the Vs and the duration of diabetes (p = 0.7534), a finding which agrees with the results published by Saglam et al. [7].

Although our pilot study comprised only 30 participants and is therefore of limited statistical power, our results compare well with previously published data on pancreas elastography. Since one very experienced examiner carried out all the measurements, we eliminated any bias from different examiners. However, the ultrasound images were not checked by a second reader, which is a definite limitation of the study. Our examiner was not blinded to the participant being a healthy volunteer or a patient. In the present study, our findings were statistically significant in the tail of the pancreas: this part of the organ is the most difficult to examine with imaging techniques and might therefore also be the most inaccurate. The precision of Vs measurement by shear wave elastography may be affected by vascular pulsation and the maximum depth of shear wave detection may be a further limitation, especially in obese patients [12].

For such a detailed analysis of changes in pancreatic morphology in T1D, factors that would limit the measurement precision, as well as variations in the examination conditions, led to individuals being excluded from the study population. Furthermore, the average duration of diabetic disease in our patients was 45 months, which is not a very long time in the course of T1D. Results could be different in patients who have had the disease for a longer period.

In conclusion, we have presented the first elastographic measurements of all pancreatic segments in patients with T1D. We found no overall difference between patients and healthy volunteers. There is only a trend indicating that chronic inflammatory processes in T1D may lead to stiffer tissue in the tail of the pancreas. Elastography is a non-invasive, reliable, and rapidly available procedure offering additional information in the diagnostic investigation of pancreatic disease. Especially for patients at risk of T1D, it might represent a promising imaging modality for screening purposes. Additional studies need to clarify whether a higher Vs is associated with T1D-specific pancreatic changes and whether this finding is related to the duration of diabetes.