Introduction Gastrointestinal stromal tumours (GISTs) are rare mesenchymal tumours of the gastrointestinal tract. In recent years there is increasing focus on immunohistochemistry biomarkers and targeted Imatinib therapy for treatment, but there is little data from the UK on factors that influence outcome.

Methods We reviewed clinical, pathological, treatment strategies, follow-up and outcome data in all patients with GISTs at our regional multidisciplinary cancer group between Jan 2008 and Dec 2012. Tumour size, mitotic index, other pathological parameters and immunohistochemical stains including CD117 (KIT), CD34, and others were recorded. Tumours were categorised according to the NIH, revised AFIP, and AJCC risk-stratification models. Cox proportional hazard regression was used to determine independent factors associated with survival.

Results 42 patients with GIST were identified. 36(85.7%) were located in the stomach, 5 (11.9%) in the small intestine, and 1(2.4%) in the oesophagus. Median age was 68 (range 43–91) yrs. 24 (57.1%) were female. Tumour size ranged from 1.0–12.7 cm (mean 5.5 cm). Metastasis was present in 19 (45.2%) at diagnosis, the liver being the most common site in 8(42.1%). Histology and immunohistochemical analysis was available in 31(73.8%). Commonest histological subtype was spindle cell in 17 (53.1%), epitheloid in 9 (29.0%) and mixed in 5 (16.1%). CD117 was positive in 90.6%, and CD34 in 75.0%. 54.8% patients underwent surgical resection with radical surgery in 47.8%, 5 of whom received extensive adjacent organ resection. 47.8%, 34.8% and 52.2% patients were categorised as high risk according to NIH, AFIP and AJCC (stage III-IV) risk models respectively. Recurrence was confirmed in 5 (11.9%) patients at a median of 597 (range 402–943) days. Of these, 2 were deemed low risk by all three classification systems. Imatinib was given to 14/42 (33.3%) patients; as primary therapy in 10 (28.3%) patients (9 palliative and 1 neoadjuvant), and as adjuvant therapy in 4 patients. Cox proportional hazard regression revealed age, tumour site, size, mitotic count, metastases at diagnosis and AFIP scores to be independently associated with worse prognosis.

Conclusion A combination of surgery and imatinib may be necessary to provide a curative treatment for GISTs and prevent recurrence. Although our study is limited by small numbers, current risk-categorisation models appear to over-estimate recurrence risk with discrepancies in predicting behaviour for certain low-risk tumours. A weighted scoring system combining independent factors associated with poor prognosis may serve as a more accurate clinical prediction tool.

Disclosure of Interest M. Friel: None Declared, Y. Vishwanath: None Declared, H. Wescott: None Declared, N. Wadd: None Declared, A. Dhar Consultant for: Honoraria for advisory board to the Pharmaceutical Industry, Speaker bureau with: Falk Pharma UK, Warner Chilcott UK, Shire Pharmaceuticals, Ferring Pharmaceuticals.