GISTs are common mesenchymal tumours and are encountered predominantly in the stomach (60%–70%), small intestine (25%–35%), and colo-rectum (10%) [3]. Only 1-5% of GISTs occur in the duodenum with the most common location being its second part [4]. Duodenal GIST with purely medial extra-luminal extension, especially from the second or third portion of the duodenum, may be difficult to diagnose pre-operatively. It may mimic non-functioning P-NET, islet cell tumour, retroperitoneal mass and PEComa which is reported in the literature, akin to our case. On reviewing the literature, we could find only 11 such cases where D-GIST masquerades as pancreatic head mass (Table 1) and the surgeon must be aware of such a unique conundrum [5–15].
The common symptomatology of all such differentials makes a pre-operative diagnosis difficult. The common presentation of D-GIST is recurrent pain in the epigastrium followed by a history of gastrointestinal bleeding due to the highly vascular duodenal submucosa. The other presentation may be palpable mass, features of gastric outlet obstruction, or infrequently diagnosed on routine screening. On the other hand, P-NETs are infrequent with an incidence of 1–2% among all pancreatic neoplasms and more than one-third of cases are non-functional. Hence, the presentation of the disease usually doesn’t provide a convincing clue towards etiology. The mean tumour size of D-GIST reported across literature was 6 cm (range 1.5–31 cm) similar to the presented case [4,16].
P-NETs, as well as D-GISTs show intense arterial enhancement on CT scan, hence may be challenging to diagnose based only on imaging. In 10-20% of cases, D-GISTs have been misdiagnosed as pancreatic head mass [16,17]. Therefore, it is suggested that in patients with large pancreatic head mass without jaundice, duodenal GIST should be kept as a differential.
Usually, the EUS is helpful in confirming the origin of D-GIST which is either from 2nd (muscularis mucosae) or 4th (muscularis propria) layer of the duodenum. However, in larger and predominantly extraluminal tumours, this may not be possible due to its compression effect, as happened in our case. Further, EUS-FNA has been reported to be effective in diagnosing such lesions with lower risk of complications. However, in context to differentiate D-GIST from pancreatic tumour, there are concerns about the post-procedure bleeding as these lesions are vascular. Further, the sensitivity of EUS-FNA cytology is better for the gastric GISTs as compared to the duodenal origin, as the reported yield is suboptimal in nearly 2/3rd of cases [2]. In our case, pancreatic tuberculosis/sarcoidosis were kept as differentials as tumour stromal cells were misinterpreted as epithelioid granulomas. The limitation in our case was that we could not perform EUS guided biopsy.
The optimal surgical procedure for D- GISTs, even with a pre-operative diagnosis, is not distinct due to the complex anatomy of the pancreaticoduodenal region. The wide local excision or segmental duodenectomy has been reported to be oncologically and technically safe, but Pancreatoduodenectomy (PD) is warranted in cases where tumour size is large, involving second or third part of the duodenum, close to the ampulla of Vater or other vital structures and in cases with a diagnostic dilemma as in our case. It is reported that more than 2/3rd cases of duodenal GISTs underwent PD due to diagnostic uncertainty and the large size of the tumour [5,11,16].
GIST of the small intestine is more aggressive and has more chances of recurrence than its gastric counterpart. Gold and colleagues have developed a nomogram based on tumour site, size, mitotic rate to assess recurrence free survival and selection of a patient for adjuvant therapy [18]. In context to our case with tumour size >5cm and mitotic rate ≤5 per 50 high power field, the risk of metastasis is noted in nearly one-fourth of the patient [3]. Hence, imatinib therapy for 3 years is beneficial to prevent future recurrence.
In conclusion, GIST located in the second part of the duodenum is seldom diagnosed correctly in the pre-operative phase especially with a lesion in the pancreatic head or mesenteric side of the duodenum. The attempt to establish preoperative diagnosis should be done with EUS guided biopsy, whenever feasible. Pancreatoduodenectomy is a safe and viable option with low morbidity for such lesions.