Search

Imaging plays an active role in the surveillance of gastrointestinal stromal tumours (GISTs). Risk stratification schemes, based on size, mitotic count, and anatomical site of origin of the GIST, help in planning preoperative and postoperative imaging strategies especially in determining the frequency and duration of surveillance; however, there is no clear consensus on the optimal imaging strategies in patients with GISTs who are completely cured by surgery and patients who are at risk of recurrence. In addition, current surveillance protocols depend on the resectability of the primary tumour and presence of metastatic disease. The objective of this article is to provide a comprehensive review of the role of the different imaging methods for surveillance of GISTs, focusing on the guidelines recommended by National Comprehensive Cancer Network and European Society of Medical Oncology - European Network for Rare adult solid Cancers, and to propose practical guidelines for surveillance of GISTs for various risk categories.

AIM To compare the diagnostic performance of prostate magnetic resonance imaging (MRI) with an endorectal coil (ERC) to performance without an ERC using either body-array (BAC) or pelvic phased-array coil (PAC) in staging T3 prostate cancer. MATERIALS AND METHODS An electronic search of the PUBMED and EMBASE databases was performed until 10 October 2018 to identify studies performing a head-to-head comparison of prostate MRI using a 1.5 or 3 T magnet with an ERC and with a BAC/PAC for staging T3 prostate cancer. Pooled sensitivity and specificity of all studies were plotted in a hierarchical summary receiver operating characteristic plot. The diagnostic performance of the two techniques in staging T3 disease was evaluated using bivariate random-effects meta-analysis. RESULTS Eight studies comparing head-to-head prostate MRI with an ERC and with a BAC/PAC were identified of which six studies compared the diagnostic performance. The pooled sensitivity and specificity of MRI with an ERC for detecting T3a, T3b and T3a+b was 53% and 95%; 52% and 92%; 72% and 65% respectively. For MRI with a BAC/PAC these were 34%, and 95%; 45% and 94%; 70% and 66%. There was no statistical difference between an ERC and a BAC/PAC in terms of sensitivity (p=0.41) and specificity (p=0.63) for T3a. The area under the receiver operating characteristic (AUROC) curve for T3a, T3b and T3a+b was 0.830, 0.901, 0.741 for an ERC and 0.790, 0.645, 0.711 for BAC, respectively. CONCLUSION There is no significant difference in the diagnostic performance of MRI of prostate with an ERC and with a BAC/PAC in staging T3 prostate cancer.

To investigate the yield of imaging in patients with relapsed prostate cancer (PC) with a low trigger prostate-specific antigen (PSA).This institutional review board (IRB)-approved, Health Insurance Portability and Accountability Act (HIPAA)-compliant retrospective study included all 133 patients (mean age 68 years; range 45-88; median 69 months since original diagnosis; interquartile range [IQR]: 32-139) with hormone-sensitive PC (HSPC, n=28) or castration-resistant PC (CRPC, n=105) and trigger PSA4 ng/ml, who underwent same-day bone scintigraphy and computed tomography (CT; total 224 time points) at Dana-Farber Cancer Institute from January to December 2013. Clinical and pathological data were obtained by manual review of the electronic medical records. All the included bone scintigraphs and CT images were reviewed by a fellowship-trained oncoradiologist to record the metastatic pattern and any clinically significant non-metastatic findings.Ninety-four of the 133 (71%) patients had metastatic disease (18/28 [64%] with HSPC, 76/105 [72%] with CRPC). Forty-one of the 133 (31%) patients developed new metastatic disease and 23/133 (17%) developed new clinically significant non-metastatic findings. The incidence of osseous, nodal, and visceral metastases, and clinically significant non-metastatic findings was similar across the HSPC and CRPC groups (p0.05 for all). Fifty-seven of the 133 (43%) patients had findings seen only at CT, of which 37 had new extra-osseous findings. Only 2/133 (2%) had findings at bone scintigraphy not seen at CT, both in areas not covered on CT.Imaging frequently demonstrated new metastatic and non-metastatic findings in patients with a low trigger PSA. CT is valuable in these patients because extra-osseous findings not visible at bone scintigraphy are frequently seen.

The purpose of this review is to familiarise radiologists with the spectrum of hepatic toxicity seen in the oncology setting, in view of the different systemic therapies used in cancer patients. Drug-induced liver injury can manifest in various forms, and anti-neoplastic agents are associated with different types of hepatotoxicity. Although chemotherapy-induced liver injury can present as hepatitis, steatosis, sinusoidal obstruction syndrome, and chronic parenchymal damages, molecular targeted therapy-associated liver toxicity ranges from mild liver function test elevation to fulminant life-threatening acute liver failure. The recent arrival of immune checkpoint inhibitors in oncology has introduced a new range of immune-related adverse events, with differing mechanisms of liver toxicity and varied imaging presentation of liver injury. High-dose chemotherapy regimens for haematopoietic stem cell transplantation are associated with sinusoidal obstruction syndrome. Management of hepatic toxicity depends on the clinical scenario, the drug in use, and the severity of the findings. In this article, we will (1) present the most common types of oncological drugs associated with hepatic toxicity and associated liver injuries; (2) illustrate imaging findings of hepatic toxicities and the possible differential diagnosis; and (3) provide a guide for management of these conditions.

There is increasing focus on intrahepatic cholangiocarcinoma (IHCC) due to its rising incidence worldwide and relatively poor prognosis, with the revised TNM classification (2009) introducing a separate staging system for IHCC for the first time. In this article, we comprehensively review the current role of the radiologist in the diagnosis and management of patients with IHCC.

Aim To study the clinical and multidetector computed tomography (MDCT) features of tumour–bowel fistula (TBF). Materials and methods Fifty-one patients (27 women; mean age 57.4 years, range 30–77years) with TBF presenting to our institution between January 2005 and February 2012 were identified retrospectively from the radiology database. MDCT images before, at, and subsequent to diagnosis of TBF were reviewed by three radiologists in consensus; clinical presentation, management, and outcome were documented from electronic medical records. Results Of 51 patients, small bowel (n = 22) was the most common site with gastrointestinal stromal tumour (GIST) being the most common sarcoma subtype (n = 10). TBF was treatment-associated (TTBF) in 40 patients [78%; 22 of whom had received molecular targeted therapy (MTT)], and spontaneous (STBF) in 11 patients (22%). Thirty-one patients (61%) were symptomatic at the time of TBF detection. TTBF was more often asymptomatic (19/40 versus 1/11; Fisher's exact test p = 0.03). In the TTBF group, 16 had a partial response, seven had stable disease, and 17 had progressive disease. Treatment was discontinued or changed to an alternative regimen in 27/40 patients, and 13/40 patients continued with the same regimen. TBF persisted in 27/33 patients (82%) who underwent CT follow-up. Thirty-one of the 51 patients were deceased at the time of analysis. Time from diagnosis of TBF to death was shorter with STBF (1.8 months) than with TTBF (6.4 months). Conclusion TBF is often associated with MTT and can be seen with treatment response or progression. TTBF is more frequently asymptomatic. TBF is usually managed conservatively by discontinuing treatment, but often persists on CT follow-up.

Gastrointestinal stromal tumour resistance to treatment with imatinib occurs due to pre-existing or acquired mutations. Computed tomography and positron-emission tomography play an essential role in prompt recognition of resistance to treatment. Primary resistance to treatment, which is encountered in the first 6 months of treatment, is associated with specific mutations. Imaging of these tumours shows no anatomical or metabolic response to treatment. Secondary resistance to treatment, which develops after an initial response, is associated with a variety of mutations acquired after the start of treatment. Imaging findings of secondary resistance are of disease progression.

There is accumulating evidence that molecular phenotyping of breast cancer determines the timing, pattern, and outcome of metastatic disease. The most clinically relevant subtypes are hormonal-positive [oestrogen and progesterone receptor (ER/PR) positive], HER2 expressing, and triple-negative breast cancers (TNBCs). ER/PR-positive breast cancers demonstrate the best prognosis; however, metastases, in particular osseous disease, may develop much later. HER2-expressing breast cancers, although aggressive, have improved outcomes due to the advent of HER2-targeted therapies, with increased risk of central nervous system (CNS) relapses later. Finally, TNBCs present in younger women, BRCA1 mutations carriers, and carry the worst overall prognosis, with high incidence of CNS metastases, especially during the first 5 years of diagnosis. It is important for radiologists to understand the nuances of these breast cancer subtypes to predict metastatic behaviours and guide possible imaging surveillance.

To study the differences in the imaging features of spread from the three cancer cell lines, namely epithelial, sarcomatoid, and lymphoid, resulting in peritoneal carcinomatosis, peritoneal sarcomatosis, and peritoneal lymphomatosis, respectively.In this institutional review board-approved Health Insurance Portability and Accountability Act (HIPAA)-compliant retrospective study, an electronic radiology database was searched to identify patients with peritoneal tumour spread who underwent CT imaging at Dana-Farber Cancer Institute, a tertiary cancer institution, between January 2011 and December 2012. Out of 1214 patients with possible peritoneal tumour spread on the radiology reports, 122 patients were included with histopathologically confirmed peritoneal disease (50 randomly selected patients with peritoneal carcinomatosis and sarcomatosis each, and all 22 patients with lymphomatosis). Two blinded, fellowship-trained radiologists in consensus reviewed the CT images in random order and recorded the imaging findings of peritoneal tumour spread. The statistical analysis was performed in two steps: the first comparing incidence of various features in each group and the second step was a pairwise analysis between each cohort.Peritoneal carcinomatosis more frequently had ascites, peritoneal thickening, and omental cake (all p ≤ 0.001). Measurable nodules were less common in peritoneal carcinomatosis (p 0.001), and when present, were ill-defined and had an irregular outline (p ≤ 0.002). Peritoneal sarcomatosis more often had discrete nodules that were well defined and had a smooth outline and less frequently had ascites, peritoneal thickening, omental caking, serosal implants, and lymphadenopathy (all p ≤ 0.005). Peritoneal lymphomatosis frequently involved the omentum and mesentery, and often had associated lymphadenopathy and splenomegaly (all p ≤ 0.002).Peritoneal carcinomatosis, sarcomatosis, and lymphomatosis have distinctive patterns on imaging, which can help the radiologists to differentiate between them.

Aim To describe the multidetector computed tomography (MDCT) features of primary, locally recurrent, and metastatic duodenal gastrointestinal stromal tumours (GISTs). Materials and methods In this institutional review board-approved, Health Insurance Portability and Accountability Act of 1996 (HIPAA)-compliant, retrospective study, 25 patients [13 men, 12 women; mean age 56 years (34–74 years)] with histopathologically confirmed duodenal GISTs seen at Dana Farber Cancer Institute and Brigham and Women's Hospital from December 1999 to October 2009 were identified. The MDCT of primary tumours in six patients and follow-up imaging in all the 25 patients was reviewed by two radiologists in consensus. Electronic medical records were reviewed to document the clinical characteristics and management. Results The mean size of the primary tumour was 3.7 cm (range 2.5–5.6 cm). Three of six primary tumours were in the second and third portions of the duodenum, one in the third portion, one in the third and fourth portions, and one in the fourth portion. Three of six of the tumours were exophytic, two were both exophytic and intraluminal, and one was intramural. The tumours were well-circumscribed, round or oval masses, with few lobulations, and were either homogeneously hyper-enhancing or heterogeneously isodense at MDCT. None of the tumours had necrosis, haemorrhage, calcification, or loco regional lymphadenopathy on imaging. Sixteen of 25 (64%) patients developed metastatic disease, the most common sites being liver (14/16; 87.5%) and peritoneum (5/16; 31%). Conclusion Duodenal GISTs are well-circumscribed, round or oval masses, and occur in the second through fourth portions of the duodenum, without lymphadenopathy or duodenal obstruction. Duodenal GISTS metastasize frequently to the liver and peritoneum.