FOCUS ON PET���MRI synergy in molecular, functional and anatomical cancer imaging Giovanni Lucignani Published online: 15 May 2008 # Springer-Verlag 2008 In the wake of PET-CT, the development of PET-MRI scanners marks another milestone in the ongoing quest for high-sensitivity and high-specificity biomedical images offering exquisite morphological detail. But why the need for PET-MRI? First, PET-CT, although well established in routine clinical practice, has two disadvantages compared with PET-MRI: it does not really allow simultaneous data acquisition, and it exposes the patient to a significant radiation dose. Second, high-resolution, very high contrast morphological imaging of soft tissues, spectroscopy, and functional imaging are all features that only MRI offers. It may be easier to appreciate the value of PET-MRI if one understands the different roles of PET and MRI, the value of their combined use (in fusion or side-by-side analysis), and the advantages of performing both scans concurrently. The strenuous efforts to develop PET-MRI scanning technology are at last yielding tangible results. A look at some of the most recently published papers in this field may help us to get these concepts into focus. PET and MRI viewed as alternatives Takeshi Mori and colleagues, of the Department of Thoracic Surgery, Graduate School of Medical Sciences, Kumamoto University, Japan, compared diffusion-weighted MRI (DWI) and PET-CT with fluorodeoxyglucose (PET- FDG) in benign/malignant discrimination of pulmonary nodules/masses [1]. They started from the premise that DWI of malignant tumours can highlight different diffusion of water molecules among tissues, a variable measurable as an apparent diffusion coefficient (ADC). PET-FDG and DWI were performed prospectively in 104 patients with 140 pulmonary nodules/masses before surgery. The quan- titative FDG uptake of each lesion was recorded as a contrast ratio of standard uptake value (SUV-CR) between the lesion and the contralateral lung. The diffusion of water molecules in each lesion, expressed as a minimum ADC (ADC-min), was established. On comparison of the diagnostic results, the receiver operating characteristics curve showed ADC-min and SUV-CR cut-off values for benign/malignant discrimination of 1.1��10 mm/s and 0.37, respectively. DWI and PET-FDG had sensitivities of 0.70 and 0.72 and specificities of 0.97 and 0.79, respectively. Although the sensitivity of the two methods was compara- ble, DWI showed significantly higher specificity than PET-FDG, giving fewer false-positive results for active inflammatory lesions (p=0.03). The ADC-min and SUV- CR values showed a significant reverse correlation (r= 0.504, p0.001). According to the authors, DWI may replace PET-FDG to distinguish malignant from benign pulmonary nodules/masses. Hiroaki Nomori, with the same group of researchers, instead compared DWI and PET-FDG in the N staging of non-small-cell lung cancer (NSCLC) [2], noting that PET- FDG, in this field, is liable to give false-positive results in the presence of concurrent lymphadenitis. The authors compared the diagnostic results of PET-FDG and DWI, used prospectively in 88 patients before surgical interven- tion for NSCLC to examine 734 lymph node stations. The Eur J Nucl Med Mol Imaging (2008) 35:1550���1553 DOI 10.1007/s00259-008-0829-7 ���Focus on...��� abridgements aim to highlight only papers published within the past year and draw extensively on the texts and summaries of the articles referenced. Less recent citations are also included when deemed useful to provide background information on the topic reviewed. G. Lucignani (*) Institute of Radiological Sciences, University of Milan, Unit of Nuclear Medicine, Hospital San Paolo, Via Antonio di Rudin��, 8, 20142 Milan, Italy e-mail: giovanni.lucignani@unimi.it

detectable size (based on diameter) of metastatic foci within lymph nodes, measured on haematoxylin and eosin-stained sections, was compared between DWI and PET-FDG and found to be 4 mm in both cases. The N staging accuracy was 0.89 with DWI (where there was less over-staging) and 0.78 with PET-FDG (p=0.012). Thirty-six of the 734 lymph node stations examined pathologically had metasta- ses. Although the two methods showed no significant difference in the diagnosis of the 36 metastatic lymph node stations, DWI more accurately characterised the 698 non- metastatic stations, giving fewer false-positive results (p= 0.002). The authors suggest that DWI could replace PET- FDG in the N staging of NSCLC. Kim and colleagues, of the Department of Radiology and Center for Imaging Science, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, investigated the capacity of MRI and PET-FDG to detect recurrent ovarian cancer [3]. Tumour recurrence was evaluated, using both techniques, in 36 patients who had previously undergone primary cytoreductive surgery for ovarian carcinoma. Retrospective analyses of the histopath- ological, clinical, and radiological follow-up findings revealed 35 sites of recurrent ovarian cancer in 22 patients: pelvic recurrence in 15, peritoneal lesions in 14, lymph node metastases in four, and abdominal wall metastases in two. PET-FDG and MRI showed patient-based sensitivity and accuracy of 73% and 91% (p0.05), and of 81% and 89% (p0.05), respectively, and lesion-based sensitivity of 66% and 86%, respectively (p0.05). The lesion-based sensitivity and accuracy of PET-FDG and MRI for peritoneal lesions were 43% and 86%, and 75% and 94%, respectively (p0.05). The authors conclude that MRI is more sensitive than PET-FDG for detecting recurrence of ovarian cancer when this consists of local pelvic recurrence and peritoneal lesions. PET and MRI viewed as complementary Kong et al., of the Department of Nuclear Medicine and PET, The Royal Marsden NHS Foundation Trust, Sutton, Surrey, UK, considered the performance of PET-FDG versus contrast-enhanced whole-body CT (ceCT) in the identification of extrahepatic disease, and versus manga- nese dipyridoxyl diphosphate liver MRI (Mn-DPDP MRI) in the evaluation of liver metastases in colorectal cancer patients who were candidates for surgery [4]. Sixty-five patients with colorectal cancer and known or suspected liver metastases underwent PET-FDG, ceCT and Mn-DPDP MRI. The results were retrospectively reviewed for evi- dence of extrahepatic disease on PET-CT and ceCT, and for the presence and number of liver metastases on PET-FDG and Mn-DPDP MRI. Metastases were confirmed by histopathology or follow-up clinical/imaging evidence of disease progression or response. PET-FDG identified unexpected extrahepatic disease (not detected on ceCT) that led to modification of the surgical strategy in 17% of the patients, in addition to three other false-positive cases. The per-patient sensitivity and specificity of both PET-FDG and Mn-DPDP MRI for liver metastases were 98% and 100%, respectively, whereas on a per-lesion basis, PET- FDG and MRI gave discordant data in 15% of scans (10/ 66). MRI correctly identified more sub-centimetre metasta- ses in eight scans. PET-FDG correctly identified more metastases in one case and confirmed disease in one equivocal MRI scan. The authors drew attention to the incremental benefit of PET-FDG over conventional ceCT in identifying extrahepatic disease in metastatic colorectal cancer, and its high sensitivity and specificity for liver metastases, suggesting that the technique should be includ- ed from the start of pre-surgical evaluation. They also observed that Mn-DPDP MRI, whose use might be guided by PET-CT, is superior for small liver metastases and thus remains a prerequisite for surgical planning in patients with confined liver metastases. Squillaci and colleagues, of the Department of Diagnos- tic and Molecular Imaging, Interventional Radiology and Radiotherapy, University of Rome, ���Tor Vergata���, Italy, assessed the accuracy of whole-body MRI (WB-MRI) with a 3-T machine and of PET-FDG for the N and M staging of patients with colorectal cancer [5]. The WB-MRI scans were evaluated by two radiologists and the PET-FDG images by one radiologist and one nuclear medicine physician. Histology and/or a clinical follow-up of 3��� 6 months served as the standard of reference. These authors found lymph node involvement in 10/20 and in 15/20 cases on WB-MRI and PET-FDG, respectively. They also found metastases in the liver (27 lesions in 15 patients and 23 lesions in 15 patients on WB-MRI and PET-FDG, respec- tively), the lungs (19 in 5 patients and 25 in 7 patients on WB-MRI and PET-FDG, respectively), and the skeleton (9 in 3 patients using both types of imaging modality). These techniques made it possible to demonstrate local recurrence at the surgery site in three patients. No brain metastases were found. According to the authors, while WB-MRI is a suitable method for examining colon cancer patients, it cannot replace PET-FDG. Schmidt and colleagues, of the University Hospitals Munich-Grosshadern, Munich, Germany and the Institute of Clinical Radiology, University Hospital Mannheim, University of Heidelberg, Germany, compared the capacity of WB-MRI (1.5 and 3T) and PET-FDG to detect tumour recurrence in 33 breast cancer patients with suspected recurrence [6]. Coronal T1w-TSE and STIR sequences, HASTE imaging of the lungs, and contrast-enhanced T1w and T2w-TSE sequences of the liver, brain and abdomen Eur J Nucl Med Mol Imaging (2008) 35:1550���1553 1551