Nuclear Medicine and Molecular Imaging • Original Research Sher et al. PET/MRI and PET/CT of Pediatric Lymphoma
Downloaded from www.ajronline.org by Univ Of California-San Diego on 02/28/16 from IP address 132.239.1.231. Copyright ARRS. For personal use only; all rights reserved
Nuclear Medicine and Molecular Imaging Original Research
Assessment of Sequential PET/MRI in Comparison With PET/CT of Pediatric Lymphoma: A Prospective Study Andrew C. Sher 1 Victor Seghers1 Michael J. Paldino1 Cristina Dodge1 Ramkumar Krishnamurthy 1 Rajesh Krishnamurthy 1 Eric M. Rohren2 Sher AC, Seghers V, Paldino MJ, et al.
Keywords: Ann Arbor staging, pediatric lymphoma, PET/CT, PET/MRI, radiation dose DOI:10.2214/AJR.15.15083 Received May 21, 2015; accepted after revision October 27, 2015. Based on a presentation at the Society for Pediatric Radiology 2015 annual meeting, Bellevue, WA. This study was supported in part by a grant from Philips Healthcare. Rajesh Krishnamurthy is a principal investigator under an institutional research agreement with Philips Healthcare MRI Systems. 1 Department of Pediatric Radiology, Texas Children’s Hospital, 6701 Fannin St, Ste 470, Houston, TX 77030. Address correspondence to A. C. Sher ([email protected]). 2 Department of Radiology, The University of Texas M. D. Anderson Cancer Center, Houston, TX.
AJR 2016; 206:623–631 0361–803X/16/2063–623 © American Roentgen Ray Society
OBJECTIVE. The objective of our study was to compare the diagnostic performance of sequential 18F-FDG PET/MRI (PET/MRI) and 18F-FDG PET/CT (PET/CT) in a pediatric cohort with lymphoma for lesion detection, lesion classification, and disease staging; quantification of FDG uptake; and radiation dose. SUBJECTS AND METHODS. For this prospective study of 25 pediatric patients with lymphoma, 40 PET/CT and PET/MRI examinations were performed after a single-injection dual-time-point imaging protocol. Lesions detected, lesion classification, Ann Arbor stage, and radiation dose were tabulated for each examination, and statistical evaluations were performed to compare the modalities. Quantification of standardized uptake values (SUVs) was performed for all lesions. All available examinations and clinical history were used as the reference standard. RESULTS. No statistically significant differences between PET/MRI and PET/CT were observed in lesion detection rates, lesion classification, or Ann Arbor staging. Fifty-four regions of focal uptake were observed on PET/MRI compared with 55 on PET/CT. Both modalities accurately classified 82% of the lesions relative to the reference standard. Disease staging based on PET/MRI was correct for 35 of the 40 studies, and disease staging based on PET/CT was correct for 35 of the 40 studies; there was substantial agreement between the modalities for disease staging (κ = 0.684; p 0.72), although PET/MRI showed systematically lower SUV measurements. PET/MRI offered an average 45% reduction in radiation dose relative to PET/CT. CONCLUSION. In a pediatric cohort with lymphoma, sequential PET/MRI showed lesion detection, lesion classification, and Ann Arbor staging comparable to PET/CT. PET/MRI quantification of FDG uptake strongly correlated with PET/CT, but the SUVs were not interchangeable. PET/MRI significantly reduced radiation exposure and is a promising new alternative in the care of pediatric lymphoma patients. he role of 18F-FDG PET (FDG PET) in the management of patients with lymphoma has progressively expanded since the FDG-avidity of non-Hodgkin lymphoma was first reported in 1987 [1]. Numerous studies have shown its value in staging both Hodgkin lymphoma and non-Hodgkin lymphoma and in monitoring treatment response in patients with Hodgkin lymphoma and in patients with non-Hodgkin lymphoma [2–4]. Although not included in the published management guidelines until 2007 [5], FDG PET was incorporated into the therapy response criteria in patients with lymphoma shortly thereafter [6]. The most recent guidelines, published in 2014, report that PET/CT is superior to CT alone for the assessment of treatment response and of remission in FDG-
T
avid lymphomas [7]. The added value of FDG PET/CT comes with a significant cost, however, in the form of ionizing radiation. This cost is particularly concerning in the pediatric population because the cumulative radiation dose from PET/CT examination may attain levels reported to potentially result in secondary malignancies [8]. With advances in technology, whole-body MRI is now technically feasible, and hybrid PET/MRI systems have been approved by the U.S. Food and Drug Administration for clinical use. Because MRI has no ionizing radiation, FDG PET/MRI has an obvious advantage over PET/CT in the form of radiation dose reduction. Outstanding soft-tissue contrast and the ability to perform functional imaging and multiparametric imaging are additional advantages of PET/MRI. In
AJR:206, March 2016 623
Sher et al. TABLE 1: Workflow Comparison No. of Examinations
Time of First PET Examination From FDG Injection (min)
Time of Second PET Examination From FDG Injection (min)
Time Between PET E xaminations (min)
PET/CT performed first
9
90 ± 17
132 ± 24
42 ± 13
PET/MRI performed first
31
71 ± 15
124 ± 21
53 ± 22
Downloaded from www.ajronline.org by Univ Of California-San Diego on 02/28/16 from IP address 132.239.1.231. Copyright ARRS. For personal use only; all rights reserved
Workflow
Note—The time data are presented as means ± SDs.
a number of oncologic diseases studied in adults, including lymphoma, PET/MRI has shown results comparable to PET/CT [9–11]. Given differences in body habitus, physiology (e.g., brown fat), and therapeutic outcomes [12], the use of 18F-FDG PET/MRI (PET/MRI) in the pediatric population mandates separate consideration. Recent work by Schäfer et al. [13] examined PET/MRI in a pediatric oncologic population and showed that lesion detection was comparable to PET/CT over a range of oncologic disease entities. However, the accuracy of PET/MRI with regard to disease staging and to the quantitation of standardized uptake values (SUVs) in pediatric lymphoma patients has not been determined to date. Therefore, the goals of the current study of pediatric patients with lymphoma were to evaluate the accuracy of 18F-FDG PET/CT (PET/CT) and PET/MRI for lesion detection, lesion classification, and disease staging; compare 18F-FDG lesion quantification on PET/MRI versus PET/CT; and calculate the magnitude of the radiation dose reduction offered by PET/MRI compared with PET/CT.
A
Subjects and Methods Patient Population
This prospective HIPAA-compliant study was approved by the local institutional review board. Written informed consent was obtained from all subjects or their legal guardians before inclusion in the study. The inclusion criteria were all patients referred for whole-body PET/CT with an established diagnosis of lymphoma for the study period from May 2013 through July 2014. The exclusion criteria were refusal to participate, scheduling conflicts, language barriers precluding informed consent by research staff, technical problems with the camera, the use of anesthesia to complete both PET/CT and PET/MRI if not needed for stand-alone PET/CT, and contraindications to MRI.
31 were performed as PET/MRI first. The FOV was from the skull vertex to the bases of feet in 18 of the 40 studies and from the skull base to the mid thighs in 22 studies. The PET images were acquired over the FOV at 120–150 seconds per bed position. The time per bed position for a given patient remained constant between the two modalities regardless of examination order to reduce potential contributors to variability between the two examinations. Subsequently, the patient was transferred to the second PET modality where a second examination was performed using the same PET acquisition parameters. All patients fasted for at least 6 hours before the studies and had a blood glucose level of less than 150 mg/dL before 18F-FDG injection.
Scanning Workflows
PET/CT Protocol
Two different workflows were used (Table 1). In both workflows, PET/CT and PET/MRI were performed sequentially of the same patient on the same day. In the first workflow, denoted as PET/CTfirst, PET/CT was performed first followed by PET/MRI. In the second workflow, denoted as PET/MRI first, PET/MRI was performed first followed by PET/CT. Of the 40 paired procedures, nine were performed as PET/CTfirst and
B
The PET/CT examinations were performed on a current-generation PET/CT scanner with timeof-flight capability (TruFlight Select PET/CT, Philips Healthcare). The CT images for attenuation correction and anatomic localization were acquired with a routine low-dose protocol (120 kV, 60–100 mAs using dose modulation, pitch of 0.813, slice thickness of 5 mm). No contrast material was administered. When PET/CT was per-
C
D
Fig. 1—9-year-old boy with stage IV Hodgkin disease. A–D, Whole-body T1-weighted spoiled gradient-echo image (A) and low-dose CT image (B) provide data for attenuation correction and anatomic localization of PET on PET/MR image (C) and PET/CT image (D).
624
AJR:206, March 2016
Downloaded from www.ajronline.org by Univ Of California-San Diego on 02/28/16 from IP address 132.239.1.231. Copyright ARRS. For personal use only; all rights reserved
PET/MRI and PET/CT of Pediatric Lymphoma formed before PET/MRI, the acquisition of PET images began 90 ± 17 (mean ± SD) minutes after tracer injection compared with 124 ± 21 minutes when PET/CT was performed after PET/MRI.
PET/MRI Protocol
PET/MRI was performed on a sequential scanner consisting of a combined current-generation PET unit with time-of-flight technology and a 3-T MRI system containing 4 × 4 × 22 mm lutetium yttrium orthosilicate crystals optically coupled to photomultiplier tubes for PET (Ingenuity TF PET/MRI, Philips Healthcare). The MR images were acquired with a slab size of 6 cm and a maximum FOV of 67.6 cm. An integrated radiofrequency body coil was used using a multistation protocol. No MRI contrast agents were administered. A 3D T1-weighted spoiled gradient-echo sequence (TE, 2.3 ms; TR, 4 ms; flip angle, 10°) was acquired while the patient was breathing freely at each bed position and was used for both attenuation correction of PET images and anatomic localization of 18F-FDG uptake. Attenuation correction was performed using an automated three-segment–based algorithm provided by the scanner manufacturer [14]. When PET/MRI was performed before PET/CT, the acquisition of PET images began 71 ± 15 minutes after tracer injection compared with 132 ± 24 minutes when PET/MRI was performed after PET/CT.
Image Analysis
Lesion detection, lesion classification, and Ann Arbor staging—PET/CT and PET/MR images that had no patient-identifying information were evaluated in consensus by two board-certified radiologists who subspecialize in nuclear radiology and had 12 and 18 years’ experience in nuclear radiology, respectively, at the time of the study. The readers were given a clinical indication for imaging of “lymphoma,” but they were otherwise blinded to all other information including other clinical history, stage of disease, and other imaging studies. For the PET/MRI studies, the MR images, FDG PET images, and fused PET/MR images were available to the readers. Because of the amount of time involved in a sequential PET/MRI study, only the attenuation-correction MRI sequence was acquired for each patient; additional MRI sequences were not routinely performed. For the PET/CT studies, the low-dose CT images, FDG PET images, and fused PET/CT images were available (Fig. 1). The Intellispace Portal software package (version 6.0, Philips Healthcare) was used for initial image analysis. The readers began the analysis by identifying the lesions on the PET images: A lesion was defined as nonphysiologic focal uptake greater than the background uptake [15, 16]. All foci of non-
physiologic uptake were classified as benign or malignant on the basis of the clinical judgment of the readers. After completing lesion detection and classification, the readers assessed each study for disease stage according to the Ann Arbor staging system [17]. After these blinded analyses, the findings from PET/CT and PET/MRI with regard to lesion detection, lesion classification, and disease stage were compared between the two modalities and with the reference standard. Quantification of standardized uptake values—A lesion was included for analysis if it was identified on either the PET/CT examination or the PET/MRI examination. The corresponding PET/CT and PET/MRI examinations of the same patient were simultaneously reviewed on a workstation (MIM 6.2, MIMVista). For each examination, the maximum SUV (SUVmax) and mean SUV (SUVmean) of all lesions were quantified. Identical volumes of interest were carefully drawn around the entire lesion on the PET/CT and PET/MRI studies; adjacent physiologic tissue and non– FDG-avid tissue were excluded from the volumes of interest. Diagnostic criteria—Comparing the diagnostic accuracy of PET/MRI with that of PET/CT requires the establishment of a reference standard. In our study, not all lesions could be evaluated by tissue sampling and histology. Therefore, the reference standard was defined as the totality of the previous and follow-up imaging studies and the subsequent clinical course [18, 19]. All patients had a minimum of 6 months of clinical follow-up after the imaging examinations for this study. Benign FDG uptake was defined as uptake that resolved on follow-up examinations without an intervention or that had associated normal morphologic findings (e.g., reactive lymphadenopathy). Similarly, a mildly avid FDG focus that remained mildly FDG-avid and stable in size on subsequent imaging while other foci decreased in FDG uptake would be classified as benign if no change in clinical treatment had occurred to address the lesion. A lesion of interest was defined as malignant if a PET-positive focus became a PET-negative focus after treatment. If no subsequent imaging examinations were available, the reference standard lesion classification was determined on the basis of subsequent clinical history and treatment.
Radiation Dosimetry
Radiation exposure from the 18F-FDG PET portions of PET/CT and PET/MRI was calculated using standard millisievert (millicurie) conversion factors from publication 80 of the International Commission on Radiological Protection [20]. For pediatric populations, 1-year-old, 5-yearold, 10-year-old, 15-year-old, and adult conversion
factors are available and were used accordingly. For a conservative estimate, the conversion factor below the patient’s true age was used when the age did not exactly match the published values. Radiation exposure from the CT portion of PET/CT was estimated using the method outlined in report 96 of the American Association of Physicists in Medicine (AAPM) [21]. Briefly, the volume CT dose index values were compiled from the dose report and multiplied by the length of the scanned region to calculate the dose-length product (DLP [in units of mGy × cm]); then, k-factors (in units of mSv/mGy × cm) from Table 3 of AAPM report 96 were used to convert DLP into effective dose. The k-factors for three anatomic regions were used: head and neck, chest, and abdomen and pelvis. The appropriate k-factor for patient age was used, and similar to the effective dose calculation from 18F-FDG PET, the k-factor below the patient’s true age was used when the age did not match the published values to allow a conservative estimate. The region scanning lengths were estimated as follows: The head and neck region started at the top of the head and ended at the thoracic inlet; the chest region constituted the region from the thoracic inlet to the dome of the liver; and the abdominopelvic region was considered the remaining area of body that was scanned. The total effective dose was calculated by adding the effective doses from the contributing PET and CT portions of the PET/CT examination.
Statistical Analysis
Demographic information and time delays are expressed as means ± SD. Agreement in lesion classification and in Ann Arbor staging between PET/CT and PET/MRI was analyzed using Cohen’s kappa statistic. A kappa value of 0–0.20 was defined as slight agreement; 0.21–0.40, fair agreement; 0.41–0.60, moderate agreement; 0.61–0.80, substantial agreement; and 0.81–1, almost-perfect agreement. Differences in sensitivity and specificity between the two modalities and the standard of reference were analyzed with the exact McNemar test. Lesion SUVs on PET/CT and PET/MRI were analyzed by Wilcoxon signed rank tests, Spearman correlation coefficients, and Bland-Altman analysis. Correlation coefficients greater than 0.7 were regarded as indicating a strong correlation and between 0.3 and 0.7 as indicating a moderate correlation. A p value
Downloaded from www.ajronline.org by Univ Of California-San Diego on 02/28/16 from IP address 132.239.1.231. Copyright ARRS. For personal use only; all rights reserved
Nuclear Medicine and Molecular Imaging Original Research
Assessment of Sequential PET/MRI in Comparison With PET/CT of Pediatric Lymphoma: A Prospective Study Andrew C. Sher 1 Victor Seghers1 Michael J. Paldino1 Cristina Dodge1 Ramkumar Krishnamurthy 1 Rajesh Krishnamurthy 1 Eric M. Rohren2 Sher AC, Seghers V, Paldino MJ, et al.
Keywords: Ann Arbor staging, pediatric lymphoma, PET/CT, PET/MRI, radiation dose DOI:10.2214/AJR.15.15083 Received May 21, 2015; accepted after revision October 27, 2015. Based on a presentation at the Society for Pediatric Radiology 2015 annual meeting, Bellevue, WA. This study was supported in part by a grant from Philips Healthcare. Rajesh Krishnamurthy is a principal investigator under an institutional research agreement with Philips Healthcare MRI Systems. 1 Department of Pediatric Radiology, Texas Children’s Hospital, 6701 Fannin St, Ste 470, Houston, TX 77030. Address correspondence to A. C. Sher ([email protected]). 2 Department of Radiology, The University of Texas M. D. Anderson Cancer Center, Houston, TX.
AJR 2016; 206:623–631 0361–803X/16/2063–623 © American Roentgen Ray Society
OBJECTIVE. The objective of our study was to compare the diagnostic performance of sequential 18F-FDG PET/MRI (PET/MRI) and 18F-FDG PET/CT (PET/CT) in a pediatric cohort with lymphoma for lesion detection, lesion classification, and disease staging; quantification of FDG uptake; and radiation dose. SUBJECTS AND METHODS. For this prospective study of 25 pediatric patients with lymphoma, 40 PET/CT and PET/MRI examinations were performed after a single-injection dual-time-point imaging protocol. Lesions detected, lesion classification, Ann Arbor stage, and radiation dose were tabulated for each examination, and statistical evaluations were performed to compare the modalities. Quantification of standardized uptake values (SUVs) was performed for all lesions. All available examinations and clinical history were used as the reference standard. RESULTS. No statistically significant differences between PET/MRI and PET/CT were observed in lesion detection rates, lesion classification, or Ann Arbor staging. Fifty-four regions of focal uptake were observed on PET/MRI compared with 55 on PET/CT. Both modalities accurately classified 82% of the lesions relative to the reference standard. Disease staging based on PET/MRI was correct for 35 of the 40 studies, and disease staging based on PET/CT was correct for 35 of the 40 studies; there was substantial agreement between the modalities for disease staging (κ = 0.684; p 0.72), although PET/MRI showed systematically lower SUV measurements. PET/MRI offered an average 45% reduction in radiation dose relative to PET/CT. CONCLUSION. In a pediatric cohort with lymphoma, sequential PET/MRI showed lesion detection, lesion classification, and Ann Arbor staging comparable to PET/CT. PET/MRI quantification of FDG uptake strongly correlated with PET/CT, but the SUVs were not interchangeable. PET/MRI significantly reduced radiation exposure and is a promising new alternative in the care of pediatric lymphoma patients. he role of 18F-FDG PET (FDG PET) in the management of patients with lymphoma has progressively expanded since the FDG-avidity of non-Hodgkin lymphoma was first reported in 1987 [1]. Numerous studies have shown its value in staging both Hodgkin lymphoma and non-Hodgkin lymphoma and in monitoring treatment response in patients with Hodgkin lymphoma and in patients with non-Hodgkin lymphoma [2–4]. Although not included in the published management guidelines until 2007 [5], FDG PET was incorporated into the therapy response criteria in patients with lymphoma shortly thereafter [6]. The most recent guidelines, published in 2014, report that PET/CT is superior to CT alone for the assessment of treatment response and of remission in FDG-
T
avid lymphomas [7]. The added value of FDG PET/CT comes with a significant cost, however, in the form of ionizing radiation. This cost is particularly concerning in the pediatric population because the cumulative radiation dose from PET/CT examination may attain levels reported to potentially result in secondary malignancies [8]. With advances in technology, whole-body MRI is now technically feasible, and hybrid PET/MRI systems have been approved by the U.S. Food and Drug Administration for clinical use. Because MRI has no ionizing radiation, FDG PET/MRI has an obvious advantage over PET/CT in the form of radiation dose reduction. Outstanding soft-tissue contrast and the ability to perform functional imaging and multiparametric imaging are additional advantages of PET/MRI. In
AJR:206, March 2016 623
Sher et al. TABLE 1: Workflow Comparison No. of Examinations
Time of First PET Examination From FDG Injection (min)
Time of Second PET Examination From FDG Injection (min)
Time Between PET E xaminations (min)
PET/CT performed first
9
90 ± 17
132 ± 24
42 ± 13
PET/MRI performed first
31
71 ± 15
124 ± 21
53 ± 22
Downloaded from www.ajronline.org by Univ Of California-San Diego on 02/28/16 from IP address 132.239.1.231. Copyright ARRS. For personal use only; all rights reserved
Workflow
Note—The time data are presented as means ± SDs.
a number of oncologic diseases studied in adults, including lymphoma, PET/MRI has shown results comparable to PET/CT [9–11]. Given differences in body habitus, physiology (e.g., brown fat), and therapeutic outcomes [12], the use of 18F-FDG PET/MRI (PET/MRI) in the pediatric population mandates separate consideration. Recent work by Schäfer et al. [13] examined PET/MRI in a pediatric oncologic population and showed that lesion detection was comparable to PET/CT over a range of oncologic disease entities. However, the accuracy of PET/MRI with regard to disease staging and to the quantitation of standardized uptake values (SUVs) in pediatric lymphoma patients has not been determined to date. Therefore, the goals of the current study of pediatric patients with lymphoma were to evaluate the accuracy of 18F-FDG PET/CT (PET/CT) and PET/MRI for lesion detection, lesion classification, and disease staging; compare 18F-FDG lesion quantification on PET/MRI versus PET/CT; and calculate the magnitude of the radiation dose reduction offered by PET/MRI compared with PET/CT.
A
Subjects and Methods Patient Population
This prospective HIPAA-compliant study was approved by the local institutional review board. Written informed consent was obtained from all subjects or their legal guardians before inclusion in the study. The inclusion criteria were all patients referred for whole-body PET/CT with an established diagnosis of lymphoma for the study period from May 2013 through July 2014. The exclusion criteria were refusal to participate, scheduling conflicts, language barriers precluding informed consent by research staff, technical problems with the camera, the use of anesthesia to complete both PET/CT and PET/MRI if not needed for stand-alone PET/CT, and contraindications to MRI.
31 were performed as PET/MRI first. The FOV was from the skull vertex to the bases of feet in 18 of the 40 studies and from the skull base to the mid thighs in 22 studies. The PET images were acquired over the FOV at 120–150 seconds per bed position. The time per bed position for a given patient remained constant between the two modalities regardless of examination order to reduce potential contributors to variability between the two examinations. Subsequently, the patient was transferred to the second PET modality where a second examination was performed using the same PET acquisition parameters. All patients fasted for at least 6 hours before the studies and had a blood glucose level of less than 150 mg/dL before 18F-FDG injection.
Scanning Workflows
PET/CT Protocol
Two different workflows were used (Table 1). In both workflows, PET/CT and PET/MRI were performed sequentially of the same patient on the same day. In the first workflow, denoted as PET/CTfirst, PET/CT was performed first followed by PET/MRI. In the second workflow, denoted as PET/MRI first, PET/MRI was performed first followed by PET/CT. Of the 40 paired procedures, nine were performed as PET/CTfirst and
B
The PET/CT examinations were performed on a current-generation PET/CT scanner with timeof-flight capability (TruFlight Select PET/CT, Philips Healthcare). The CT images for attenuation correction and anatomic localization were acquired with a routine low-dose protocol (120 kV, 60–100 mAs using dose modulation, pitch of 0.813, slice thickness of 5 mm). No contrast material was administered. When PET/CT was per-
C
D
Fig. 1—9-year-old boy with stage IV Hodgkin disease. A–D, Whole-body T1-weighted spoiled gradient-echo image (A) and low-dose CT image (B) provide data for attenuation correction and anatomic localization of PET on PET/MR image (C) and PET/CT image (D).
624
AJR:206, March 2016
Downloaded from www.ajronline.org by Univ Of California-San Diego on 02/28/16 from IP address 132.239.1.231. Copyright ARRS. For personal use only; all rights reserved
PET/MRI and PET/CT of Pediatric Lymphoma formed before PET/MRI, the acquisition of PET images began 90 ± 17 (mean ± SD) minutes after tracer injection compared with 124 ± 21 minutes when PET/CT was performed after PET/MRI.
PET/MRI Protocol
PET/MRI was performed on a sequential scanner consisting of a combined current-generation PET unit with time-of-flight technology and a 3-T MRI system containing 4 × 4 × 22 mm lutetium yttrium orthosilicate crystals optically coupled to photomultiplier tubes for PET (Ingenuity TF PET/MRI, Philips Healthcare). The MR images were acquired with a slab size of 6 cm and a maximum FOV of 67.6 cm. An integrated radiofrequency body coil was used using a multistation protocol. No MRI contrast agents were administered. A 3D T1-weighted spoiled gradient-echo sequence (TE, 2.3 ms; TR, 4 ms; flip angle, 10°) was acquired while the patient was breathing freely at each bed position and was used for both attenuation correction of PET images and anatomic localization of 18F-FDG uptake. Attenuation correction was performed using an automated three-segment–based algorithm provided by the scanner manufacturer [14]. When PET/MRI was performed before PET/CT, the acquisition of PET images began 71 ± 15 minutes after tracer injection compared with 132 ± 24 minutes when PET/MRI was performed after PET/CT.
Image Analysis
Lesion detection, lesion classification, and Ann Arbor staging—PET/CT and PET/MR images that had no patient-identifying information were evaluated in consensus by two board-certified radiologists who subspecialize in nuclear radiology and had 12 and 18 years’ experience in nuclear radiology, respectively, at the time of the study. The readers were given a clinical indication for imaging of “lymphoma,” but they were otherwise blinded to all other information including other clinical history, stage of disease, and other imaging studies. For the PET/MRI studies, the MR images, FDG PET images, and fused PET/MR images were available to the readers. Because of the amount of time involved in a sequential PET/MRI study, only the attenuation-correction MRI sequence was acquired for each patient; additional MRI sequences were not routinely performed. For the PET/CT studies, the low-dose CT images, FDG PET images, and fused PET/CT images were available (Fig. 1). The Intellispace Portal software package (version 6.0, Philips Healthcare) was used for initial image analysis. The readers began the analysis by identifying the lesions on the PET images: A lesion was defined as nonphysiologic focal uptake greater than the background uptake [15, 16]. All foci of non-
physiologic uptake were classified as benign or malignant on the basis of the clinical judgment of the readers. After completing lesion detection and classification, the readers assessed each study for disease stage according to the Ann Arbor staging system [17]. After these blinded analyses, the findings from PET/CT and PET/MRI with regard to lesion detection, lesion classification, and disease stage were compared between the two modalities and with the reference standard. Quantification of standardized uptake values—A lesion was included for analysis if it was identified on either the PET/CT examination or the PET/MRI examination. The corresponding PET/CT and PET/MRI examinations of the same patient were simultaneously reviewed on a workstation (MIM 6.2, MIMVista). For each examination, the maximum SUV (SUVmax) and mean SUV (SUVmean) of all lesions were quantified. Identical volumes of interest were carefully drawn around the entire lesion on the PET/CT and PET/MRI studies; adjacent physiologic tissue and non– FDG-avid tissue were excluded from the volumes of interest. Diagnostic criteria—Comparing the diagnostic accuracy of PET/MRI with that of PET/CT requires the establishment of a reference standard. In our study, not all lesions could be evaluated by tissue sampling and histology. Therefore, the reference standard was defined as the totality of the previous and follow-up imaging studies and the subsequent clinical course [18, 19]. All patients had a minimum of 6 months of clinical follow-up after the imaging examinations for this study. Benign FDG uptake was defined as uptake that resolved on follow-up examinations without an intervention or that had associated normal morphologic findings (e.g., reactive lymphadenopathy). Similarly, a mildly avid FDG focus that remained mildly FDG-avid and stable in size on subsequent imaging while other foci decreased in FDG uptake would be classified as benign if no change in clinical treatment had occurred to address the lesion. A lesion of interest was defined as malignant if a PET-positive focus became a PET-negative focus after treatment. If no subsequent imaging examinations were available, the reference standard lesion classification was determined on the basis of subsequent clinical history and treatment.
Radiation Dosimetry
Radiation exposure from the 18F-FDG PET portions of PET/CT and PET/MRI was calculated using standard millisievert (millicurie) conversion factors from publication 80 of the International Commission on Radiological Protection [20]. For pediatric populations, 1-year-old, 5-yearold, 10-year-old, 15-year-old, and adult conversion
factors are available and were used accordingly. For a conservative estimate, the conversion factor below the patient’s true age was used when the age did not exactly match the published values. Radiation exposure from the CT portion of PET/CT was estimated using the method outlined in report 96 of the American Association of Physicists in Medicine (AAPM) [21]. Briefly, the volume CT dose index values were compiled from the dose report and multiplied by the length of the scanned region to calculate the dose-length product (DLP [in units of mGy × cm]); then, k-factors (in units of mSv/mGy × cm) from Table 3 of AAPM report 96 were used to convert DLP into effective dose. The k-factors for three anatomic regions were used: head and neck, chest, and abdomen and pelvis. The appropriate k-factor for patient age was used, and similar to the effective dose calculation from 18F-FDG PET, the k-factor below the patient’s true age was used when the age did not match the published values to allow a conservative estimate. The region scanning lengths were estimated as follows: The head and neck region started at the top of the head and ended at the thoracic inlet; the chest region constituted the region from the thoracic inlet to the dome of the liver; and the abdominopelvic region was considered the remaining area of body that was scanned. The total effective dose was calculated by adding the effective doses from the contributing PET and CT portions of the PET/CT examination.
Statistical Analysis
Demographic information and time delays are expressed as means ± SD. Agreement in lesion classification and in Ann Arbor staging between PET/CT and PET/MRI was analyzed using Cohen’s kappa statistic. A kappa value of 0–0.20 was defined as slight agreement; 0.21–0.40, fair agreement; 0.41–0.60, moderate agreement; 0.61–0.80, substantial agreement; and 0.81–1, almost-perfect agreement. Differences in sensitivity and specificity between the two modalities and the standard of reference were analyzed with the exact McNemar test. Lesion SUVs on PET/CT and PET/MRI were analyzed by Wilcoxon signed rank tests, Spearman correlation coefficients, and Bland-Altman analysis. Correlation coefficients greater than 0.7 were regarded as indicating a strong correlation and between 0.3 and 0.7 as indicating a moderate correlation. A p value
