Introduction

Morphological and molecular imaging, such as magnetic resonance imaging (MRI), computed tomography (CT), and positron emission tomography (PET), play a key role in clinical oncology: diagnosis as well as staging and treatment evaluation rely on imaging findings. However, each of these imaging procedures suffers from well-known inherent limitations. The objective of combining anatomical and functional information into a single image has long been pursued. In the wake of research efforts focusing on software coregistration, the introduction of hybrid PET/CT systems [1] substantially changed cancer imaging. Since the introduction of commercial PET/CT scanners, both PET and CT technology have continued to evolve, improving image quality [2–4]. 2-(18F)-fluoro-2-deoxy-d-glucose (FDG) PET/CT, and to a lesser extent other radiotracers, have had a major impact on oncology over the past decade, improving clinical management [5–10]. Despite the success of PET/CT, the possibility of replacing CT with MRI has been increasingly investigated [11–13], and was actually investigated even before the introduction of PET/CT [14]. These efforts are driven by a series of considerations:

  • the possibility of lower patient exposure to radiation, given that MRI does not use ionizing radiation; this aspect is important in pediatric patients as well as in other situations such as cases requiring repeated studies;

  • MRI provides high-resolution anatomical and structural images offering better soft-tissue contrast resolution and a large variety of tissue contrasts, compared with CT;

  • MRI, through functional MRI (fMRI) and MR spectroscopy (MRS), provides a wealth of additional information able to enhance the diagnostic performance and quantitative capabilities of PET and help improve patient management and understanding of tumor biology.

On the other hand, MRI has several disadvantages:

  • examination time can be substantially longer than with CT, depending on the imaging protocol;

  • there still exist contraindications related to metal implants and pacemakers; metallic implants produce a local signal loss in MRI images, misleading the image segmentation procedure and resulting in classification of the region as air instead of tissue in the MRI-based attenuation map; this may result in underestimation of uptake;

  • a relatively low sensitivity for lung lesions, although new scanning parameters have dealt with this limitation; moreover, one or both lungs are occasionally not identified correctly and instead interpreted as air, thus leading to undercorrection of standardized uptake values in this area;

  • MRI is more expensive than CT.

In contrast to CT, MRI offers a multitude of “endogenous” contrasts and a high capability for soft-tissue differentiation, as well as many exogenous contrast media ranging from gadolinium-based agents to highly specific cellular markers. Having sensitivity in the picomolar range, PET is ideally suited for the visualization of specific molecules in living organisms. However, PET lacks the spatial resolution offered by MRI, which in turn lacks sensitivity. Therefore, the combination of PET and MRI is highly complementary. Unlike CT, MRI results in no additional radiation. Furthermore, as mentioned previously, MRI has interesting functional imaging capabilities, such as perfusion, BOLD effect, diffusion, and spectroscopy.

However, MRI cannot simply replace the CT component of a PET/CT scanner, since a fully integrated PET/MRI combination requires technological modifications of both the PET and the MRI. Indeed, MRI interferes with state-of-the art PET technology (i.e., PMT) and PET interferes with field gradients or MR radio frequency. The physics of and technical solutions to these problems are beyond the scope of this review and many excellent papers dealing with these issues have been published [15–18]. Moreover, practical PET/MRI implementations should also take in account the issue of accurate MRI-based methods for attenuation correction (AC) of the measured PET emission data, which is particularly important for quantitative PET. Whereas CT measures the attenuation coefficient of tissues at X-ray energies, the MRI signal is determined by tissue hydrogen density and relaxation properties, and thus, the derivation of 511-keV-photon attenuation coefficients is much more complex than in CT AC. Different methods for deriving attenuation maps from MR have been proposed [19–23].

Following the installation of the first head-only PET/MRI scanners [24], whole-body PET/MRI scanners have now been introduced in the market for clinical use. Taking into account the technological challenges inherent in combining PET and MRI, the three major companies (Siemens Healthcare, Philips Healthcare, and GE Healthcare) have proposed different PET/MRI designs.

As installation of the first PET/MRI systems is ongoing and only a few are already operating, the challenge is to understand the clinical potential of this new modality. Although the published data on integrated PET/MRI is still limited, first experiences with this imaging technique consistently indicate a better performance in those indications requiring high soft-tissue contrast. Furthermore, the true value of this new hybrid imaging modality lies in the simultaneous acquisition of fMRI data and metabolic PET information.

The aim of this review is to underline the oncological clinical applications of PET/MRI on the basis of published data on hybrid PET/MRI, when available, or on PET/CT and MRI. The current status of and clinical indications for PET/MRI in oncology will be discussed. A first section of the paper deals with whole-body FDG PET/MRI in cancer staging (TNM) and therapy response assessment. Thereafter, the focus shifts to non-FDG tracers and evaluation of tumor biology. Finally, a short paragraph is devoted to several methodological considerations.

Whole-body FDG PET/MRI

T staging

FDG PET/CT has been shown to be capable of accurately determining T stage in a few indications, such as head and neck tumors, non-small lung cancer and colorectal cancer [25–28]. However, since the evaluation of local tumor extent relies mainly on morphological data because of high spatial resolution, and the functional component of hybrid imaging does not add information on T stage, PET/MRI could prove to be superior to PET/CT in those indications in which MRI has been found to be more accurate than CT. On this basis, it can be anticipated that PET/MRI would be suitable in all those indications in which the high tissue contrast of MRI would allow higher accuracy: breast, head and neck, liver, musculoskeletal, and brain tumors.

PET/MRI performed better than PET/CT in head and neck cancer [29], but did not add significant information to MRI alone in either simultaneous or fused imaging [29, 30]. In our own unpublished experience, MR/PET showed better soft-tissue contrast resolution (Fig. 1).

Fig. 1
figure 1

A 67-year-old male patient with squamous cell carcinoma of the upper alveolar ridge. From left to right, upper row: T1 TSE tra FAT SAT (after contrast), T2 TSE tra and PET/MRI fusion (FDG); lower row: DWI (b = 800), CT and CT (bone window). PET/MRI allows better lesion identification and anatomical location compared with PET/CT. In addition bone marrow involvement was found allowing correct T3 staging

Full size image

In breast cancer, MR mammography shows high sensitivity and relatively low specificity, while FDG PET/CT is more specific and less sensitive [31, 32]. An increase in specificity from 53 to 97 % was observed by adding FDG and MRI data [33]. However, a study with software-based fusion of images did not show any significant benefit [34]. This discrepancy could be due to lesion dimensions, since small lesions are known to show poor FDG uptake [32].

FDG PET/CT has a low sensitivity in hepatic primary tumors (hepatocellular cancer), with radiotracers other than FDG showing better results [35]. No data on PET/MRI in this setting have been published so far; however, it is conceivable that combined PET/MRI, with the addition of fMRI, would perform as well as MRI alone.

T staging of colorectal cancer with software-based PET/MRI fusion is inaccurate [36], and it is improbable that a hybrid PET/MRI system would perform better than the MRI alone or even FDG PET/CT [28].

Although MRI, including fMRI, is the method of choice in T staging of primary bone tumors [37], high accuracy using FDG PET/CT has been reported [38]. The high soft-tissue contrast of MRI has made it the preferred modality in soft-tissue sarcomas with no advantage derived from adding FDG data [39, 40]. However, in both instances, i.e., the primary bone tumors and soft-tissue sarcomas, FDG PET/MRI, providing FDG uptake information, may be used to guide biopsies and in preoperative and radiation treatment planning [41].

The literature on FDG PET/CT in the staging of both Hodgkin and non-Hodgkin lymphomas is extensive, showing very high accuracy, which has led to a widespread use of this methodology [42–44]. Diffusion-weighted whole-body MRI has been proposed as an alternative to FDG PET/CT, showing comparable accuracy [45, 46]. However, more recent data suggest a more conservative approach, since overstaging was observed [47]. PET/MRI could be an alternative to PET/CT in lymphoma, particularly in younger patients, although no published data are available to date.

MRI offers an alternative morphological imaging approach avoiding additional radiation exposure in pediatric oncology patients. Thus, in children, PET/MRI can add significant information to that obtained through the use of separate imaging modalities. Data on the diagnostic value of combined and registered image analysis of FDG PET and MRI for staging and re-staging in pediatric oncology have recently been published [48]. On a lesion-based analysis, PET/MRI performed significantly better than the two imaging modalities alone [48]. Moreover, in our experience, appropriate MRI scanning parameters allow PET/MRI imaging of the lung (Fig. 2).

Fig. 2
figure 2

A 77-year-old male patient with suspected lung nodule. a Whole-body coronal MRI (T2 STIR), FDG PET, and integrated PET/MRI. b Axial MRI (T2 HASTE and DWI), FDG PET, and PET/MRI

Full size image

N staging

FDG PET/CT is more accurate than CT alone in the assessment of N stage in different malignant diseases, and shows significantly higher accuracy than whole-body MRI in staging lymph node metastatic spread [49, 50]. This advantage of FDG PET/CT over MRI and CT is due to the metabolic data: PET has the capability to detect viable tumor tissue in metastatic lymph nodes independently of their size. Instead, MRI relies on size-based lymph node assessment in N staging. However, both benign and malignant lymph nodes vary significantly in size: nodes larger than 1 can be benign, just as those smaller than 1 cm can be malignant. Therefore, PET/MRI is expected to perform as well as PET/CT, or even better by adding fMRI data. The major limitation would be, as in PET/CT, identification of small metastases or micrometastases because of spatial resolution. Available data from the literature are still limited, but initial reports are encouraging. A higher accuracy of PET/MRI compared with PET/CT has been observed in head and neck squamous cell cancer, although 30 % of patients were understaged [30]. Figure 3 shows a patient from our experience. No major advantage of PET/MRI over MRI alone would be expected in lymph node staging in breast cancer, where the clinical scenario must be established also taking into account other modalities such as sentinel lymph node biopsy. FDG PET/CT has been reported to show an accuracy of 80 % in N staging of colorectal cancer [28], comparable to MRI [51], while combined FDG PET/CT and MRI showed higher values [52]. FDG PET/CT has proven to be superior to all conventional imaging modalities, including MRI, in soft-tissue sarcomas [53], and again PET/MRI can be expected to perform at least as well, or better by integrating fMRI data. Analogously to breast cancer, the sentinel lymph node plays a major role in melanoma depending on tumor thickness. However, when the lymph node biopsy is positive imaging is recommended [54], and integrated PET/MRI will surely be an invaluable tool in melanoma patients.

Fig. 3
figure 3

A 72-year-old female patient with squamous cell carcinoma of the tongue base and lymph node involvement. From left to right, top row: STIR coronal (transaxial view), PET/MRI fusion, PET; second row: DWI (b = 800), ADC map, CT; third row: CT, STIR coronal, DWI (b = 800); bottom row: STIR coronal, PET/MRI fusion. PET/MRI characterized both lesion extent and nodal involvement better than CT/PET did

Full size image

M staging

Some studies have compared FDG PET/CT with whole-body MRI for evaluation of distant metastases [6, 50, 55–57]. Results are different in different anatomical districts, with PET/CT found to be more accurate in assessing lung metastases and MRI in liver, bone, and obviously, brain metastases. However, certain sequences (single-shot, turbo spin echo) as well as diffusion-weighted imaging can lead to improvement in lung metastasis detection [58, 59]. PET/MRI might be expected to perform better than MRI in the same regions. Liver MRI, especially with the use of specific contrast agents, is remarkably more accurate than FDG PET/CT in detecting hepatic metastases [60]. The sensitivity of Gd-EOB-DTPA-enhanced MRI and retrospectively fused PET/MRI in the detection of liver metastases has been shown to be higher than that of PET/CT, while PET/MRI resulted in a non-significant increase in sensitivity when compared to contrast-enhanced MRI [61]. Although the first available clinical application of integrated PET/MRI was the brain, initial studies focused on primary brain tumors [24, 29]. Data from brain evaluation using whole-body MRI protocols demonstrated higher accuracy, compared with FDG PET/CT, in the detection of cerebral metastases [58]. However, whether the diagnostic performance of integrated FDG PET/MRI would be of clinical relevance is difficult to predict from the available information. Actually, different acquisition procedures could potentially benefit PET/MRI: dual-phase FDG imaging, dynamic contrast-enhanced MRI, and spectroscopy. Detection of bone metastases is based on imaging: X-ray skeletal survey, bone scintigraphy, CT, MRI and PET/CT. In a meta-analysis comparing data from more than 15,000 patients, MRI and FDG PET/CT were found to be equally accurate [62]. Different results have been found in studies comparing the two imaging modalities in different cancer types, with MRI showing better sensitivity in the detection of bone metastases in breast cancer and lower sensitivity in lung cancer patients [63, 64]. Moreover, diffuse bone marrow infiltration is difficult to detect on PET/CT [57], but more reliably detected on MRI. Thus, for detecting bone metastases, integrated PET/MRI has the potential to overcome the limitations of each modality and become the procedure of choice.

Therapy response assessment

Although therapy response in solid tumors has usually been based solely on systematic assessments of tumor size, using, for example, the WHO, RECIST, and International Workshop Criteria (IWC) for lymphoma [65–67], these methods do have well-recognized limitations. For this reason, there has been a considerable and growing interest in molecular imaging techniques. FDG PET/CT has been shown to improve response assessment in several tumor types ranging from malignant lymphoma to a variety of solid tumors [67, 68]. In recent years, clinical studies have demonstrated that mid-treatment diffusion-weighted imaging could be used as an imaging response biomarker [69–72]. In the context of integrated PET/MRI, the addition of fMRI would enhance the performance of both modalities in the evaluation of both treatment response and recurrences.

A first prospective study of PET/MRI in therapy response assessment of head and neck tumors showed excellent results, combining the high negative predictive value of PET with the high sensitivity of MRI [30] (Fig. 4). In breast cancer, MRI and fMRI have proven to be sensitive to early therapy response [73], and FDG PET/CT is capable of predicting early response to neoadjuvant therapy [74]. The potential of integrated PET/MRI must be fully explored, since this hybrid methodology will likely play a key role in evaluating recurrence and therapy response in breast cancer patients. Both in chemoembolization and in radio frequency ablation of liver tumors, both primary and metastatic, FDG PET/CT and MRI (with the addition of fMRI) demonstrated better performances than CT, with high accuracy [75, 76]. Similar results have been obtained in the evaluation of neoadjuvant chemoradiotherapy in colorectal cancer [77, 78]. Although FDG PET/CT monitoring of therapy of lymphomas (Hodgkin and diffuse large B cell) has been included in the guidelines [42], more recent studies reported low and maybe suboptimal predictive value [79, 80] in early therapy assessment, probably related to therapy schemes different from classic chemotherapy, with higher predictive value in end-therapy evaluation. In this setting, MRI with the addition of fMRI (especially diffusion-weighted imaging) showed excellent and complementary results [81]. Additional advantages to be derived from integrated PET/MRI in the therapy response assessment of lymphomas would be the possibility of differentiating recurrence from thymic rebound by the use of chemical shift MRI [82] and the possibility of using the modality in children, thanks to the reduction of radiation exposure as compared with PET/CT. Figure 5 shows a patient with Hodgkin lymphoma from our experience.

Fig. 4
figure 4

A 47-year-old male patient with squamous cell carcinoma of the rhinopharynx treated with chemotherapy, radiotherapy and right latero-cervical lymphadenectomy. From left to right, top row: T1 TSE FAT SAT tra (post contrast), PET/MRI fusion, PET; middle row: ADC map, DWI (b = 800), CT; bottom row (coronal view): T1 TSE FAT SAT (after contrast), PET/MRI fusion, DWI (b = 800), PET. PET/MRI leads to correct identification of nodal relapse

Full size image
Fig. 5
figure 5

A 56-year-old female with Hodgkin lymphoma. a Whole-body coronal MRI (T2 STIR), FDG PET, and integrated PET/MRI. b Axial MRI (T2 HASTE and DWI), FDG PET, and PET/MRI. Abdominal lymph node involvement is present and a suspicious breast nodule is found

Full size image

Non-FDG tracers

FDG is utilized in more than 90 % of cancers in staging, re-staging, assessing therapy response and during the follow-up. However, not all tumors show a significant increase of metabolic activity on FDG imaging. Non-FDG tracers already used for clinical applications are 11C- and 18F-choline, 11C-methionine and 18F-FET, 18F-DOPA, 68Ga-DOTA-somatostatine analogs, 11C-acetate, 18F-FLT, and 124I (Table I). Imaging of intracranial masses was performed on a hybrid PET/MRI system capable of simultaneous PET/MRI acquisition administering 11C-methionine in patients with gliomas and 68 Ga-DOTA-somatostatine analogs in those with meningiomas [83]. Similar diagnostic image quality on the hybrid PET/MRI and the PET/CT studies was found, with excellent agreement between semi-quantitative data (i.e., the tumor-to-reference tissue ratios) of the PET/MRI and PET/CT systems (r = 0.98) [82]. 18F-FDOPA PET/MRI fusion provides precise anatomical localization of tracer uptake in patients with gliomas, with PET data capable of identifying tumors not visible on MRI [84]. In high-risk differentiated thyroid cancer patients, fused 124-iodine PET/MRI has proven to be superior to 124-iodine PET/CT in detecting lesions, particularly those <10 mm [85]. Both modalities were performed after thyroidectomy but before radioiodine therapy, and data provided by fused PET/MRI imaging provided more precise and tailored dosimetry [85]. MRI and fMRI (spectroscopy) is used to localize, stage, and obtain functional information in prostate cancer patients [86]. On the other hand, current evidence indicates that PET/CT with 11C-acetate and 11C-choline or 18F-fluorocholine might be useful in the diagnosis and staging of primary tumors, in guiding tumor biopsy, in detection of metastatic disease, in monitoring of response to therapies, and in prognostication [87]. Integrated PET/MRI imaging would be of clinical benefit by adding the diagnostic capabilities of each modality. Parametric fusion PET/MRI based on 11C-choline PET/CT and diffusion-weighted MRI for the identification of primary prostate cancer improves identification when compared to each modality alone [88]. A significant increase in accuracy, mainly due to a remarkable improvement in specificity, was found when fused 11C-acetate PET/MRI was compared with the performance of each individual modality alone in the detection of prostate cancer [89].

Table 1 Non-FDG PET tracers
Full size table

Although the majority of data published so far deals with fusion PET/MRI, there is an increasing evidence that integration of the two modalities would yield better performances in the setting of non-FDG tracers.

Tumor biology

Both in the preclinical and in clinical oncology applications there is a growing need for in vivo detection techniques to obtain better characterization of cellular and subcellular processes. Molecular imaging as stand-alone or multimodality methods has become an important discipline. Among the latter PET/CT has emerged as a successful imaging method and the recent introduction of PET/MRI will pave the way for a better understanding of oncologic processes, in both the clinical and preclinical settings. PET provides information on metabolic and molecular parameters with high sensitivity of data at molecular level, but limitations as regards tissue morphology. On the other hand, MRI provides a variety of data from the anatomical to the metabolic level, the latter with lower sensitivity at the molecular level but to a large extent synergic to those of PET. Several aspects of the biology of cancer might be evaluated with integrated PET/MRI: from metabolism to angiogenesis, from proliferation to apoptosis, from stem cells to metastases. Each of these could provide further insights into the biology and pathophysiology of cancer. More importantly from a clinical point of view, molecular imaging can be efficiently used as imaging surrogate in monitoring therapies and in prognosis evaluation. At present, only PET is able to evaluate cellular metabolism and proliferation in vivo, by using appropriate tracers such as FDG, l-DOPA, methionine analogs, and FLT (Table 1) [90–92]. Both PET, with 18F-fluoromisonidazole, and MRI, with spectroscopy and BOLD fMRI, can be used to study hypoxia [90–92]. Data on apoptosis can be obtained from PET imaging with radiolabeled annexin or indirectly from diffusion-weighted MRI [90–92]. Dynamic contrast-enhanced MRI and PET with perfusion tracer are used to assess the tumor perfusion. In addition, molecular markers such as integrins, VEGF and its receptors, and metalloproteinase can be evaluated using appropriate labeled radiotracers or MRI with paramagnetic nanoparticles [90–92].

PET/MRI, in which each modality provides supplementary information to that provided by the other, is destined to play a major role as a surrogate-imaging marker. Moreover, the introduction of new MRI probes and PET tracers should further extend the clinical application of this hybrid methodology, allowing a better understanding of cancer biology.

Methodological considerations

Technical and methodological aspects of PET/MRI are addressed more extensively in another article, in this issue [93]. Certainly, to fully explore the clinical applications of this imaging modality, several methodological problems should be addressed. Briefly, the quality of PET/MR imaging, particularly of PET imaging, should be comparable to that obtained with PET/CT, including anatomical localization and semi-quantitative data. In addition, in oncology, optimization of the scan protocol, workflow and image analysis is more complex for integrated PET/MRI than for clinical PET/CT.

In fused PET and MRI imaging, the results of Dixon-based MR imaging, a water/fat separation technique used for MR-based attenuation correction for PET/MRI, were evaluated for anatomical correlation of PET-positive lesions on a 3 T clinical scanner compared to low-dose CT [94]. No significant difference was found in anatomical localization for all PET-positive lesions in low-dose CT compared to Dixon-based MR [94]. Moreover, SUV obtained from PET/MRI showed a high concordance with SUV from PET/CT, 6.31 ± 4.52 vs 6.36 ± 4.47 [94]. Fully integrated PET/MRI showed more than encouraging results in patients with oncologic diagnoses with excellent correspondence with PET/CT in terms of image quality, contrast and alignment [95]. No significant difference was found in detection rate both on a lesion and on a patient basis, and quantitative evaluation showed high correlation between SUVs obtained by PET/MRI and PET/CT both in lesions and in background tissue [93]. A high concordance (ranging from 88 to 99 %) between PET/MRI and PET/CT on both a patient and a lesion basis was observed by other authors [96, 97].

Thus, integrated clinical whole-body PET/MRI is feasible, producing images of high quality. Despite different attenuation approaches, qualitative and quantitative image analysis showed excellent correlation with PET/CT.

Quantitative PET is playing an increasing role in oncology, and different approaches have been used to quantify tracer uptake by PET/CT: from full kinetic analysis to SUV (SUVmean, SUVmax, SUVpeak), metabolic tumor volume and total lesion glycolysis. To date few papers have directly compared semi-quantitative SUV measurements determined using PET/MRI versus PET/CT [95, 96]. Both SUVmax and SUVmean determined by PET/MRI were significantly lower than the corresponding values obtained by PET/CT [95, 96]. This difference was observed in lesions as well as in background, and in specific organs (i.e., liver, lung, bone, spleen, and muscle), and ranged from −11 % in lesions to −33 % in bone and −56 % in liver [95, 96]. Differences in time interval between tracer administration and scan acquisition cannot explain the lower SUV values observed with PET/MRI. Actually, the PET/MRI scans were started later than (56–88 min after) the PET/CT imaging [95, 96]. The fact that FDG uptake expressed as SUV is well known to increase over time in malignant lesions means that, the finding of significantly lower values in lesions detected with PET/MRI is related to other factors. On the other hand, FDG uptake in normal tissue decreases over time, but this decrease is lower than the differences observed in both these studies [95, 96]. Thus, these findings suggest an underestimation of semi-quantitative assessment of tracer uptake in PET/MRI. Many factors influence SUV quantification, and among them attenuation correction may play a major role in PET/MRI. It has recently been reported that increasing the numbers of tissue classes used to produce attenuation maps substantially reduces the relative error in measuring SUV leading to a significant decrease in underestimation of SUV by PET/MRI in both lesions and normal tissue [98, 99]. Further studies are needed to address these issues, and even more importantly, to assess the reproducibility of quantitative data obtained with PET/MRI.

Since functional and molecular data can be obtained using MRI, the design of clinical protocols for whole-body PET/MRI can be more complex than it is for PET/CT. In addition, decisions should be taken on complete or partial body coverage as well as on specific MRI acquisition protocols. Actually, patients can be referred for PET/MRI with prior imaging results, including PET/CT, or without such results. In the first scenario, a partial PET/MRI scan can be applied, with a few bed positions, and a more complete MRI examination, including fMRI. Conversely, when no or few prior imaging results are available, a whole-body PET/MRI is mandatory, followed by specific MRI sequences to address specific questions raised by the whole-body scan. It should be underlined that each choice will have an impact on imaging time, workflow and image analysis. Moreover, while no particular artifacts are to be expected in MRI images, MRI-based attenuation correction would produce specific artifacts on PET images due to metallic implants, truncation, particular tissue effects (i.e., bone and lung), and the misregistration that is occasionally present in integrated PET/MRI. A first paper dealing with workflow and scan protocols in oncology with a specific integrated PET/MRI scanner has been published [100], and similar efforts for other commercial scanners are expected in the near future.

Conclusions

PET/MRI has the potential to repeat the success of PET/CT in clinical oncology, particularly in those indications, such as soft-tissue tumors, where MRI offers advantages over CT. The data obtained from fMRI appears complementary to the information provided by PET, either using FDG or non-FDG tracers. Although the literature on integrated PET/MRI in oncology is limited, the first experiences with this technique as well as data from fused PET and MRI imaging support its diagnostic utility. Moreover, a clear advantage of PET/MRI over PET/CT is the lower exposure of patients to radiation. PET/CT delivers a relatively high-absorbed radiation dose, compared with the dose from a regular chest radiograph. Since MRI does not use any ionizing radiation, it can conceivably be used without restrictions in serial studies, in children, and in all these situations in which radiation exposure could be a concern. Thus, it is conceivable that PET/MRI could replace PET/CT in all these settings.

However, at a time when the value of PET/CT is still questioned by many health administrators on the basis of cost concerns, the development and the subsequent clinical introduction of PET/MRI, a more expensive and more complex imaging modality, will pose new and major challenges for physicians, industries and administrators. Scientific literature, able to give answers to specific questions in particular clinical settings, is awaited before routine clinical oncology applications of PET/MRI can be proposed. Whether PET/MRI will replace PET/CT or be a complementary methodology depends on what value is added by PET/MRI to cancer imaging, showing the modality to constitute true progress. Moreover, cost-efficiency issues will play a major role in the decision-making process. On the basis of current knowledge, it can be anticipated that hybrid PET/MRI will probably not replace PET/CT in all cases, but rather be used as an integration in some applications and as a stand-alone hybrid technology in other indications and in children.