Search

Aim To evaluate volume change thresholds for reliable volumetric evaluation of pulmonary metastatic tumour response in comparison with the response evaluation criteria of solid tumours (RECIST). Methods Fifty consecutive patients with pulmonary metastases undergoing follow-up chest multidetector-row CT under chemotherapy were prospectively included. Metastatic volumes were estimated twice by two independent observers using commercially available automated volumetry software. Intra- and interobserver agreements of metastatic volume change were estimated by 95% limits of agreement. The response to chemotherapy, as defined by an incrementally increasing percentual metastatic volume change (20%–70%), was assessed by extended Kappa statistics. Categorical agreement was correlated with percentual volume change thresholds as defined for pulmonary tumour response. Results A total of 202 metastases were evaluated. The 95% intra- and interobserver limits of agreement were −2.16%–1.88% and −1.79%–2.14%, respectively. Partial response, total remission or progressive disease were present in 46% of patients using a 70% volume change threshold and in 48%, 56%, 62%, 70% and 82% of patients using 60%, 50%, 40%, 30% and 20% thresholds, respectively. Discordant response ratings occurred in 0%–6% of patients. General combined categorical agreement on treatment response was very good (κ=0.94–1), but was diminished with volume thresholds 0.05). Kappa values correlated in linear fashion with the threshold values (r2=0.868; p, Aim Relapse after chemoradiation for non-small cell lung cancer (NSCLC) carries a dismal prognosis. A prospective positron emission tomography (PET)-response database was used to estimate how many patients could benefit from response-adapted surgery. Method Pre- and post-treatment fluorodeoxyglucose (FDG)-PET scans were performed for 88 patients who received 60 Gy with concurrent platinum-based chemotherapy (n=73) or 60 Gy alone (n=15). Response categories were complete metabolic response (CMR), partial metabolic response (PMR), stable metabolic disease (SMD), or progressive metabolic disease (PMD). We considered that medically operable patients who attained a PMR with potentially resectable residual disease confined to the lung and patients who attained a CMR but subsequently relapsed locally with anatomically resectable disease could benefit from surgery. Twenty-four medically inoperable patients could not have tolerated thoracotomy, (predominantly chronic obstructive airways disease, but also heart disease (n=6), rheumatoid arthritis (n=1), and renal failure (n=1)). Of the remaining 64 patients, 29 attained CMR (45%), 25 PMR (39%), 4 SMD (6%) and 6 PMD (9%). Of 29 CMR patients, 5 ultimately had local relapses that could potentially have been prevented by surgery, 15 never relapsed and 9 relapsed with extensive disease. Of 26 PMR patients, 12 might have benefited from salvage surgery as they had absent (n=10) or minimal (2) evidence of mediastinal disease and potentially resectable residual parenchymal disease. Of 64 patients, potentially fit for thoracotomy, 17 might have benefited from salvage or adjuvant surgery after chemoradiation. To benefit 5 CMR patients, surgery would have been necessary in 29, including 15 who may have already been cured. PMR patients represent the best candidates for response-adapted salvage surgery., Aim To estimate the diagnostic accuracy of dynamic contrast enhanced (DCE) computed tomography (CT) and magnetic resonance (MR) imaging, and [18F]fluorodeoxyglucose (FDG) positron emission tomography (PET) and Tc-99m depretiode single photon emission computed tomography (SPECT) imaging for the evaluation of solitary pulmonary nodules using a meta-analysis. Data sources Studies published between January 1990 and December 2005 in PubMed. Study selection Studies that examined (DCE) CT and MR imaging patterns and MR imaging signal characteristics, and studies that examined [18F]FDG PET, Tc-99m depretiode SPECT imaging for the evaluation of solitary pulmonary nodules and; enrolled at least 10 participants with a pulmonary nodule, with histological confirmation of the majority of lesions; and presented sufficient data to permit calculation of contingency tables were included in the analysis. Data extraction Two reviewers independently abstracted data regarding true positive, false positive, false negative and true negative of the imaging test, and independently assessed study quality and disagreements were resolved by a third reviewer. Data synthesis Forty-five studies met the inclusion criteria in total, 10 DCE CT studies, 6 DCE MRI studies, 22 [18F]FDG PET studies, and 7 99mTc-depreotide SPECT studies. Study methodological quality was fair. Sample sizes were small, minimum 20, maximum 356 and mean 65 subjects. We used a meta-analytic method to construct summary receiver-operating characteristic curves. These showed no statistically significant difference between the modalities although DCE CT performed slightly less well. Conclusions Dynamic contrast-enhanced CT and MR imaging and FGD PET and 99mTc-depreotide SPECT imaging are accurate non-invasive imaging tests for distinguishing malignant from benign solitary pulmonary nodules or mass lesions., Aim To assess, separately and in combination, the diagnostic impact of high resolution computed tomography (HRCT) and 99mTc-depreotide (a somatostatin receptor scintigraphy), in patients with pulmonary lesions verified by CT. Methods We included 127 patients presenting with one or more pulmonary lesions suspected of malignancy on CT. Supplementary HRCT and 99mTc-depreotide tomography were performed. HRCT findings were classified into three groups: high, intermediate and low probability of malignancy. The final diagnosis was based on cytology, histology or CT follow-up. Results Sixty-one out of 127 patients had lung cancer. The overall sensitivity, specificity and accuracy of 99mTc-depreotide were 95%, 58%, and 76%, respectively. The sensitivity of 99mTc-depreotide in tumours >15 mm was 100%; the three false negatives were tumours ≤15 mm. HRCT showed high probability of malignancy in 58 patients; 48 had lung cancer. Twenty patients were classified as low probability, one was false negative. The intermediate group consisted of 49 patients (39%); 12 had lung cancer. When adding the intermediate group to the high probability group sensitivity, specificity and accuracy of HRCT were 98%, 29%, and 62%, respectively. Of the 49 patients in the intermediate group 99mTc-depreotide was positive in 23 patients (nine true positive) and negative in 26 (three false negative). Conclusion The sensitivity of 99mTc-depreotide alone was high in patients with lung lesions >15 mm. HRCT alone was accurate in the group of patients with high and low probability of malignancy. However, HRCT was indeterminate in 39% of the patients. In this group additional 99mTc-depreotide contributed to further diagnostic strategy., Aim Accurate target volume delineation is essential for effective radiotherapy. Pelvic lymph nodes are not easily identifiable on conventional imaging, but can be visualised by contrast-enhanced magnetic resonance imaging (MRI) using intravenous ultra-small particles of iron-oxide (USPIO). We have previously reported pelvic node clinical target volume (CTV) delineation guidelines for use with conventional imaging, based on nodal mapping studies using USPIO. The aim of this study was to independently verify these guidelines in a further cohort of patients. Methods Ten patients with gynaecological malignancy underwent MRI with and without intravenous USPIO. MR sequences were transferred to a radiotherapy 3D planning system. The proposed guidelines were used to outline a pelvic node CTV on pre-contrast T2-weighted images. On post-contrast T2 *-weighted images the pelvic nodes were identified and outlined. The pre- and post-contrast images were co-registered and CTV examined for node coverage. Results By applying the guidelines, full coverage of 737 node outlines out of a total of 741 was achieved. Of the 4 outlines not fully covered, 2 were in the anterior external iliac and 2 in the lateral external iliac node regions. Conclusion MRI using USPIO has enabled the production of guidelines for localising a pelvic node CTV with conventional imaging. Application of these guidelines to a further cohort of patients resulted in coverage of 99.5% of node outlines demonstrating the reliability of this technique., Aim To record water diffusion coefficients of normal uterine structures and to correlate water diffusion with tumour grade and stage of endometrial cancer. Methods Seventeen women with endometrial cancer (85% undergoing hysterectomy) were evaluated at 1.5 and 3 T using multiplanar T2-weighted, dynamic T1-weighted contrast enhanced and single shot echo-planar imaging (EPI) diffusion-weighted magnetic resonance imaging (MRI) (b=0, 50, 150, 500, 1000 s /mm 2) for staging purposes. Apparent diffusion coefficients (ADC) (from all b-values) of imaging determined and histologically confirmed normal myometrium, cervix stroma and cervical mucosa were obtained. Tumour ADCs were correlated with histological grade and pathological stage. Results An overlap in ADC values between cervical mucosa (median = 1.33 × 10 −3 mm 2/s; inter-quartile range = 0.95–1.52 × 10 −3 mm 2/s) and myometrium (median = 1.25 × 10 −3 mm 2/s; inter-quartile range = 0.97–1.41 × 10 −3 mm 2/s) was observed but both were greater than cervix stroma (median = 0.62 × 10 −3 mm 2/s; inter-quartile range = 0.23–0.82 × 10 −3 mm 2/s) (Mann–Whitney test; p=0.002). Fifteen tumours, 13 endometrioid carcinomas (grade 1=6; grade 2=3; grade 3=6), 1 clear cell (grade = 3) and 1 mullerian tumour (grade = 3) were found at pathology. The pathological tumour stages for the 17 tumours were T0 = 1, T1a/b = 2, T1c = 8, T2 = 3, T3 = 3. There was a significant difference between the ADC values of tumours (median = 0.73 × 10 −3 mm 2/s; interquartile range = 0.66–0.78 × 10 −3 mm 2/s) and myometrium (p=0.001) with greater restriction in tumours. There was a negative Kendall’s rank correlation between adenocarcinoma ADC and histological grade (τ=−0.52, p=0.018) and pathological stage (τ=−0.60, and p=0.006); that is aggressive lesions had lower ADC values. Conclusion Intrinsic differences in water diffusion profile between normal myometrium and tumour enable tumour detection on MRI and also reflect aggressiveness of histological lesion., Aim To evaluate and compare the performance of [123I]meta-iodobenzylguanidine (MIBG) with cross sectional imaging in the detection and localisation of adrenal and extra-adrenal phaeochromocytomas. Methods All patients between 1993 and 2005 with adrenal or extra-adrenal phaeochromocytomas that had computed tomography (CT) and magnetic resonance (MR) imaging were identified. Patients included in the study had biochemical or clinical evidence of a phaeochromocytoma or were part of a paraganglioma syndrome. Imaging features on CT and MRI including tumour location, size, and CT contrast enhancement, MRI signal characteristics and MIBG positivity were compared. In particular tumours not detected on MIBG were closely interrogated. Results We identified 65 patients with 72 adrenal phaeochromocytomas, 1 of which had metastases at presentation. 60 of these had [123I]MIBG (9 negative/51 positive). We identified 38 patients with extra-adrenal tumours (18 neck, 2 chest, 17 abdomen, 4 pelvis, 2 spine) including 7 with disseminated metastases and either resected primary or unknown primary at presentation. Thirty-three of these had MIBG studies (12 negative/21 positive). The sensitivity of MIBG when compared against CT and MRI was 85% and 63% for adrenal and extra-adrenal tumours respectively. The radiological features of phaeochromocytomas in patients with MIBG-negative studies are emphasised. Conclusion MIBG is more sensitive in detecting phaeochromocytomas at adrenal compared to extra-adrenal locations. MIBG is more likely to be negative in adrenal tumours with only a thin rim of peripheral viable tissue and central necrosis. No specific feature predicts negativity for extra-adrenal sites., Aim Malignant melanoma is the commonest cause of skin-cancer related deaths worldwide. Nodular melanoma is an aggressive type of melanoma due to its vertical growth and tendency to metastasise. Its incidence may also be increasing. The aims of this study were to review a cohort of patients with histologically proven nodular melanoma to determine if the incidence is increasing. We also investigated the incidence and distribution of metastatic spread of nodular melanoma based on follow-up computed tomography (CT) imaging. Methods A retrospective review was performed using a local database of melanoma patients in a tertiary referral centre. Radiological investigations were reviewed and evidence of the development of visceral, skeletal or soft tissue metastases based on follow up CT imaging was recorded. Results A total of 269 patients with malignant melanoma were identified. Histological analysis was available in 144 of these patients. Nodular melanoma accounted for 46% (n=66) of these patients. Metastatic disease was identified in 15% (n=10) of patients with nodular melanoma, either at presentation or on follow-up CT examination. The commonest sites of metastatic spread included lung (50%), liver (30%), spleen (20%), skin (20%) and peritoneum (20%). Sentinel lymph node mapping was negative in 50% (n=5) of patients who subsequently developed distant metastases. Conclusion Nodular melanoma is an aggressive condition and may be increasing in incidence with 46% prevalence in our study. Sentinel node status is not a reliable predictor of disease progression in patients with nodular melanoma. Nodular melanoma metastasised to distant sites more often than other histological subtypes in our study and therefore close follow-up of these patients is recommended., Aim To discuss the role of contrast-enhanced ultrasonography (CEUS) in the evaluation of soft-tissue tumors and to correlate these results with those from computed tomography (CT) and magnetic resonance imaging (MRI). Methods Twenty-seven patients with soft tissue tumors evaluated with SonoVue contrast agent (Bracco, Italy) and Technos software (Esaote, Italy) were reviewed. Cases included: malignant fibrous histiocytoma (2), myxoid liposarcoma (2), well-differentiated liposarcoma of the salivary gland (1), recurring fibrosarcoma (2), Merckel tumour (1), breast MALToma (1), low-grade sarcoma (1), hepatocellular carcinoma (HCC) subcutaneous seeding (2), soft-tissue metastasis (5), benign nerve sheath tumor (2), post-traumatic neuroma (2), lipoma (3), intramuscular haemangioma (1), elastofibroma dorsi (2). We subjectively evaluated intralesional microcirculation and macrocirculation, intensity and distribution of contrast enhancement and identified five patterns: (1) absence of enhancement; (2) slight enhancement; (3) diffuse macrocirculation with slow wash-out; (4) macrocirculation and microcirculation with fast enhancement and quick wash-out; (5) diffuse, intense, and heterogeneous enhancement with macrocirculation and persistent microcirculation. Ten of these patients were examined with contrast enhanced CT, 17 patients with MRI. Results Pattern 1 was found in lipomas, pattern 2 in elastofibroma dorsi and post-traumatic neuroma, pattern 3 in haemangioma and neurinomas, pattern 4 in fibrosarcoma, Merckel tumor, well-differentiated sarcoma and seeding HCCs; pattern 5 in malignant fibrous histiocytomas, myxoid liposarcomas, and subcutaneous metastasis. Only in the well-differentiated liposarcoma of the salivary gland, did CEUS demonstrate a pattern 1 with erroneous diagnosis of lipoma. In all other lesions there was good correlation with CT, MRI, and histological exams. Conclusion CEUS is a non-invasive diagnostic method in the initial work-up of soft-tissue tumors. The description of vascular patterns can be important for diagnosis and differentiation of benign from malignant tumours.

Aim To evaluate diffusion-weighted magnetic resonance imaging (DW-MRI) for detection of regional nodal metastases in primary head and neck squamous cell carcinoma (HNSCC). Materials and methods Ten patients with HNSCC underwent an MRI prior to surgery. The scan protocol consisted of plain transverse T1- and T2-weighted turbo spin echo (TSE) sequences and gadolinium-enhanced transverse, coronal and sagittal T1-weighted TSE sequences. Additionally, a transverse echo-planar DW-MRI sequence, using six b-values between 0 and 1000s/mm2, was performed. Apparent diffusion coefficient (ADC) maps were calculated for the entire b-value range (ADC avg) and for the high b-values separately (b≥500s/mm2, ADC high). ADC values were compared with histopathology of the neck dissection specimens. After determining an optimal threshold for ADC avg and ADC high, sensitivity and specificity was calculated. Finally DW-MRI was compared with TSE-MRI for detection of nodal metastases. Results MRI showed 55 lymph nodes; 41 were benign and 14 malignant on histopathology. The mean ADC avg value for benign lymph nodes was 0.00122±0.00029mm2/s and 0.00090±0.00014mm2/s for malignant nodes (p, Aim To investigate the incidence, localization and prognostic influence of retropharyngeal (RP) nodal involvement in patients with squamous cell carcinoma of the oropharynx. Methods The CT studies of 208 patients presenting with oropharyngeal carcinoma were retrospectively analysed. The location of nodal neck disease was registered according to recent consensus guidelines for target volume delineation, and special attention was given to the RP nodes. The influence on recurrence (local and regional), distant metastasis, and survival was investigated. Results RP adenopathies were present in 16% (34/208) of all patients and in 23% (31/134) of patients with nodal disease in other neck sites. A solitary ipsilateral RP node was present in 3 (9%) of 34 patients with RP nodes; 2 of these 3 patients had a primary posterior pharyngeal wall tumour. No patients presented with a solitary contralateral RP node. At 5 years, patients with RP adenopathy had significantly more regional recurrences (45% vs. 10%, p=0.004). Involvement of RP lymph nodes significantly (RR 4.29 [95% CI 3.33–5.25], p=0.01) and independently predicted regional recurrence in multivariate analysis. Disease-specific survival was significantly lower in the RP node positive group (38% vs. 58%, p=0.03). Conclusions Given the high incidence of RP nodal involvement in oropharyngeal cancer and the negative impact on prognosis, RP nodes should be included in the target volume, especially in node-positive necks. In node-negative necks, inclusion of RP nodes into the target volume is advised in posterior pharyngeal wall tumours., Aim Acanthomatous change (AC) has been described as a sign of greater malignancy. However, it is generally accepted that histomorphological variations have no bearing on biologic behaviour or prognosis and that there is no correlation between histologic subtype and radiologic/imaging findings. Materials and methods In a review of 20 cases of aggressive ameloblastoma over a 12-year period (1991–2002), AC was often stated in histological reports, without any special clinical significance or pathological quantification, radiological correlation or proven clinical implication. In this study, AC was classified as focal or diffuse (from 1+ to 4+) and correlated with the radiological appearance and clinical behaviour. Results AC occurred in the follicular and plexiform subtypes. Fourteen cases were recurrences. It was conclusively shown that progressive AC was associated with an aggressive radiological appearance and biologic behaviour, including a case of ameloblastic carcinoma. Conclusion AC correlates with aggressive behaviour. Pathologists and clinicians should be alerted to the significance of AC in biopsy and resected histological specimens. These cases should be carefully assessed by MRI for diagnosis and follow up., Aim To investigate sentinel lymph node mapping in patients with early stage breast carcinoma. Methods A systematic search of world literature of sentinel node mapping in patients with early stage breast carcinoma following chemotherapy was undertaken. Using the keywords ‘sentinel node biopsy’, ‘early stage breast cancer’, ‘axillary lymph node dissection’ and ‘lymphatic mapping’, potentially eligible studies were identified. Results Between 2000 and 2005, 24 trials were reported that met eligibility criteria. Of the 1247 patients studied, 1047 patients (84%) had successfully mapped lymph nodes. The proportion of patients who had successful lymph node mapping ranged from 63% to 100%, with 55% of studies reporting a rate of, Aim To investigate the behaviour of the dynamic optical breast imaging (DOBI) curve in relation to the microvessel density (MVD) count of surgical specimens from breast biopsies. Methods Forty-six patients underwent DOBI evaluation for mammographic findings suggesting biopsy. The DOBI evaluation was performed the day before or on the day of the scheduled biopsy. The MVD count was performed from the site of the specimen where the pathology was located. The characteristics of the DOBI curve were correlated to the MVD count and to the pathology results of the biopsy. Results All malignant lesions had a high MVD count and a DOBI curve with a downslope direction, rather straight and without any initiation delays. The benign cases with a high MVD count had a downslope DOBI curve but not always straight and with some initiation delays. The rest of the cases had a low MVD count and most of them an elevated DOBI curve. Conclusion Our preliminary results indicate a relationship between malignant breast lesions with a characteristic DOBI curve and high MVD count., Aim To assess magnetic resonance (MR) imaging features of breast lymphoma (BL) and to compare MR characteristics of BL and ductal infiltrating carcinoma (DIC). Methods MR features of 7 patients with 7 breast lymphoma lesions (BL) and 7 patients with 9 DIC lesions were retrospectively evaluated by two radiologists, according to a multifactorial evaluation protocol (8-point scale) for lesion morphology (form, margins, and enhancing pattern) and dynamics (initial wash-in peak and post-initial washout). Signal intensity values of pre-contrast, initial and post-initial enhancement were recorded in order to calculate percentage changes of these values for BL and DIC lesions. Diameter of lesions was also recorded. The Wilcoxon two-sample test was used to evaluate the differences between BL and DIC MR features. The same statistical test was used to assess the inter-observer variability. Results BL showed a significantly higher median diameter (median 38 mm, range 25–50 mm) than DIC (median 20 mm, range 15–23 mm) (p0.05). There was no significant inter-observer variability for all the assessed data (p>0.05). Conclusion MR evaluation of BL and DIC lesions did not show significant difference in initial enhancement percentage change and overall evaluation score. There was significant difference between the two groups in diameter and percentage change of post-initial enhancement., Aim To determine the presence, severity and distribution of morphological changes consistent with ‘fibrosis’ and fatty infiltration on MRI with SPIO in patients with colorectal metastases, and their relationship to chemotherapy. Methods One-hundred and seventeen patients with colorectal metastases, 89 undergoing chemotherapy and 28 untreated, underwent unenhanced in-phase and opposed-phase T1-weighted gradient refocused echo (GRE) imaging, and T2-weighted fast spin echo (FSE) and GRE sequences before and after 8 μmol /kg ferucarbotran. The images were reviewed by two experienced observers to determine the presence, severity and distribution of morphological changes consistent with ‘fibrosis’ and fatty infiltration. Percentage signal intensity loss (PSIL) and tumour-to-liver contrast (CNR) were measured on superparamagnetic iron oxide (SPIO)-enhanced liver images. Results In the post-chemotherapy patients the mean PSIL, mean CNR were 51.9, 6.7 on FSE and 70.7, 29.6 on GRE sequences, respectively. In untreated patients the mean PSIL, mean CNR were 54.9, 8.3 on FSE and 74.1, 32.4 on GRE. Twenty-five patients (19 treated, 6 untreated) had ‘fibrosis’ (4 severe, 21 moderate), which was extensive in 11 and localised in 14. Compared with the non-fibrotic group (n=92), the 25 patients with ‘fibrosis’ had significantly lower PSIL (47.5 v 54.1 [p, Aim To monitor by CTp perfusion, changes in hepatocellular carcinoma (HCC) following therapy with thalidomide. To correlate tumour perfusion with alpha-fetoprotein (AFP) and circulating endothelial cells (CEC). To investigate whether baseline tumour perfusion may predict therapy outcome. Methods Thirteen consecutive patients with advanced HCC, treated with orally administered thalidomide (200 mg /day), underwent baseline CTp and follow-up CTps, every 8 weeks. On May 2006 nine patients reached follow-up >6 months. Perfusion parameters (blood flow (BF), blood volume (BV), mean transit time (MTT), permeability surface (PS)) of the tumours were calculated using dedicated software (CT Perfusion 3, GE) and statistically correlated with AFP and CEC. Results Twelve patients were assessable for response (one excluded for toxicity); none had partial or complete remission; 6 had stable disease (SD) at 6 months and 3 had progressive disease (PD). All patients with PD showed increased BF (average 30.57%) at the time of PD; all patients with SD had stable or reduced BF (average −13.3%) at 6 months. Significant correlation between BF and AFP changes was found (R=0.71). Baseline BF and BV of HCC were significantly higher (p, Aim Pseudomyxoma peritonei (PP) is a rare condition characterised by mucinous ascites and mucinous implants involving the peritoneal surfaces. The pathology of PP has been defined into three groups: disseminated peritoneal adenomucinosis (DPAM), peritoneal mucinous carcinomatosis (PMCA) and the hybrid type. The aim of the retrospective study was to describe the computed tomography (CT) findings of 51 patients referred to our institution for surgical cytoreduction. Methods Two observers independently reviewed the CT images performed between 1996 and 2004. Results There were 38 patients with PMCA (18 women, 20 men, mean age 59 years) and 13 with DPAM (8 women, 5 men, mean age 55 years). No hybrid type of PP was encountered. Pleural effusion or pleural rind, omental cake and paraaortic lymph nodes were seen only in cases of PMCA. Liver and spleen scalloping as well as hernias were also common in PMCA. Calcifications were seen in 36% of DPAM and 26% of PMCA examinations. Conclusions Pseudomyxoma peritonei is difficult to diagnose clinically. However, knowledge of typical CT findings can help the radiologist to suggest a diagnosis of PP., Aim To compare prospectively the accuracy of ferucarbotran-enhanced high-resolution dynamic T1-weighted imaging and delayed T2-weighted gradient refocused echo (GRE) imaging for detecting small metastases. Methods We studied 25 patients referred for liver resection with colorectal metastases. All were imaged with dynamic T1-weighted fat-suppressed 3D GRE (VIBE) and delayed T2-weighted GRE sequences after bolus injection of ferucarbotran. Images were independently evaluated by two blinded observers who identified and localised lesions using a four-point confidence scale. Only lesions smaller than 2 cm were included in the analysis. The results were correlated with surgery, intra-operative ultrasound and histopathology. AFROC analysis was used to determine the accuracy of each technique. Results One-hundred and forty-seven metastases, smaller than 2 cm (107 < 1 cm) were present. For all metastases the accuracy values were 0.94 and 0.92 for T1-weighted and 0.93 and 0.91 for T2-weighted, respectively, for each of two observers. Accuracy values for lesions of 1 cm or less were 0.93 and 0.89 for T1-weighted and 0.91 and 0.88 for T2-weighted. Nine sub-centimetre lesions were not detected by either observer on any technique. Overall 11/147 (7.5%) lesions were detected only on T1-weighted (mean of two observers) compared with 7/147 (4.7%) detected only on T2-weighted (all less than 1 cm). Conclusion Accuracy for ferucarbotran-enhanced T1- and T2-weighted imaging was similar. With combined T1- and T2-weighted imaging we detected substantially more lesions than either sequence alone., Aim To compare the performance of tri-phasic multidetector computed tomography (MDCT) and gadoxetate (formerly known as Gd-EOB-DTPA, Primovist ®, Schering, Germany)-enhanced magnetic resonance (MR) imaging in detection and characterization of focal liver lesions. Methods The study population consisted of 34 patients, who underwent triphasic MDCT and Gd-EOB-DTPA enhanced 1.5-T MRI (time interval, max 14 days), including dynamic and late phase gadoxetate-T1-weighted scans at 20 and 40 min. Two radiologists evaluated the images in consensus reading with respect to lesion size, number, location and characterisation. Gold standard was defined by histopathologic correlation, intra-operative ultrasound or imaging follow up. The McNemar test was used for statistical analysis. Results Of a total of 126 lesions CT detected 78% and MRI 93% (p=0.003), respectively. The correct characterisation of lesions was accomplished in 64% with CT and in 89% with MRI (p, Aim To evaluate the preoperative TNM-staging of oesophageal carcinomas with contrast enhanced 16-or 64-row multidetector computed tomography (MDCT) and to compare the findings with results by endoscopic ultrasound (EUS), using the histopathological findings as gold standard. Methods Eighty-six patients with oesophageal carcinoma (proven by endoscopic biopsy) were preoperatively examined with MDCT in a prone position. After distending the oesophagus (gas granules) and stomach (1.5 L of water) the chest and neck were subjected to MDCT scanning with 16×0.75 mm or 64×0.63 mm, the abdomen with 16×1.5 or 64×1 mm, using an individualised contrast injection protocol based on a bolus tracking technique. EUS was performed using a fibre-optic endoscope with a 5–10 MHz electronic array with a 360 degree scanning angle. Our MDCT staging criteria were drawn from a careful review of the literature and from personal experience and we used the TNM classification of the American Joint Committee on Cancer. Results Both modalities detected all tumours with a sensitivity of 100%. T staging by MDCT and EUS reached accuracies of 79% and 89%. The diagnosis of local node involvement by MDCT was correct in 85%, in EUS in 72%. The evaluation of distant node involvement and other metastases reached an accuracy of 92% with MDCT and only 65% with EUS. Conclusion Non-invasive MDCT is an important tool in the preoperative staging of oesophageal carcinoma and is complementary to EUS.