Prospective evaluation of planar bone scintigraphy, SPECT, SPECT/CT, 18 F-NaF PET/CT

Transcription

1 Acta Oncologica ISSN: X (Print) X (Online) Journal homepage: Prospective evaluation of planar bone scintigraphy, SPECT, SPECT/CT, 18 F-NaF PET/CT and whole body 1.5T MRI, including DWI, for the detection of bone metastases in high risk breast and prostate cancer patients: SKELETA clinical trial Ivan Jambor, Anna Kuisma, Susan Ramadan, Riikka Huovinen, Minna Sandell, Sami Kajander, Jukka Kemppainen, Esa Kauppila, Joakim Auren, Harri Merisaari, Jani Saunavaara, Tommi Noponen, Heikki Minn, Hannu J. Aronen & Marko Seppänen To cite this article: Ivan Jambor, Anna Kuisma, Susan Ramadan, Riikka Huovinen, Minna Sandell, Sami Kajander, Jukka Kemppainen, Esa Kauppila, Joakim Auren, Harri Merisaari, Jani Saunavaara, Tommi Noponen, Heikki Minn, Hannu J. Aronen & Marko Seppänen (2016) Prospective evaluation of planar bone scintigraphy, SPECT, SPECT/CT, 18 F-NaF PET/CT and whole body 1.5T MRI, including DWI, for the detection of bone metastases in high risk breast and prostate cancer patients: SKELETA clinical trial, Acta Oncologica, 55:1, 59-67, DOI: / X To link to this article: Published online: 02 Apr Submit your article to this journal Article views: 1818 View related articles View Crossmark data Citing articles: 28 View citing articles Full Terms & Conditions of access and use can be found at Download by: [ ] Date: 22 December 2017, At: 08:53

2 Acta Oncologica, 2016; 55: 59 67= = ORIGINAL ARTICLE Prospective evaluation of planar bone scintigraphy, SPECT, SPECT/CT, 18 F-NaF PET/CT and whole body 1.5T MRI, including DWI, for the detection of bone metastases in high risk breast and prostate cancer patients: SKELETA clinical trial Ivan Jambor 1, Anna Kuisma 2, Susan Ramadan 2, Riikka Huovinen 2, Minna Sandell 3, Sami Kajander 4, Jukka Kemppainen 4, Esa Kauppila 5, Joakim Auren 3, Harri Merisaari 4, Jani Saunavaara 3, Tommi Noponen 6, Heikki minn 2,4, Hannu J. Aronen 1,3 & Marko Seppänen 4,6 1 Department of Diagnostic Radiology, University of Turku, Turku, Finland, 2 Department of Oncology and Radiotherapy, University of Turku, Turku, Finland, 3 Medical Imaging Centre of Southwest Finland, Turku University Hospital, Turku, Finland, 4 Turku PET Centre, Turku, Finland, 5 Department of Clinical Physiology and Nuclear Medicine, North-Karelia Central Hospital, Joensuu, Finland and 6 Department of Clinical Physiology and Nuclear Medicine, Turku University Hospital, Turku, Finland Abstract Purpose. Detection of bone metastases in breast and prostate cancer patients remains a major clinical challenge. The aim of the current trial was to compare the diagnostic accuracy of 99m Tc-hydroxymethane diphosphonate ( 99m Tc-HDP) planar bone scintigraphy (BS), 99m Tc-HDP SPECT, 99m Tc-HDP SPECT/CT, 18 F-NaF PET/CT and whole body 1.5 Tesla magnetic resonance imaging (MRI), including diffusion weighted imaging, (wbmri DWI) for the detection of bone metastases in high risk breast and prostate cancer patients. Material and methods. Twenty-six breast and 27 prostate cancer patients at high risk of bone metastases underwent 99m Tc-HDP BS, 99m Tc-HDP SPECT, 99m Tc-HDP SPECT/CT, 18 F-NaF PET/CT and wbmri DWI. Five independent reviewers interpreted each individual modality without the knowledge of other imaging findings. The final metastatic status was based on the consensus reading, clinical and imaging follow-up (minimal and maximal follow-up time was 6, and 32 months, respectively). The bone findings were compared on patient-, region-, and lesion-level. Results. 99m Tc-HDP BS was false negative in four patients. In the region-based analysis, sensitivity values for 99m Tc-HDP BS, 99m Tc-HDP SPECT, 99m Tc-HDP SPECT/CT, 18 F-NaF PET/CT, and wbmri DWI were 62%, 74%, 85%, 93%, and 91%, respectively. The number of equivocal findings for 99m Tc-HDP BS, 99m Tc-HDP SPECT, 99m Tc-HDP SPECT/CT, 18 F-NaF PET/CT and wbmri DWI was 50, 44, 5, 6, and 4, respectively. Conclusion. wbmri DWI showed similar diagnostic accuracy to 18 F-NaF PET/CT and outperformed 99m Tc-HDP SPECT/CT, and 99m Tc-HDP BS. Breast cancer (BCa) and prostate cancer (PCa) are currently the most common cancer types in the western world [1]. Despite of advances in local and adjuvant treatment approximately 15 20% of patients will relapse and 70% of these will later develop bone metastases [2]. The patients with bone metastases are not amenable to curative therapy but will need hormonal and/or chemotherapy with non-steroid antianalgesics, opioids, bisphophonates and radiotherapy reserved for palliation [3]. The recognition of skeletal disease as an important metastatic site has elicited new forms of targeted therapies such as inhibition of receptor activator of nuclear factor-kb (RANK) ligand and radionuclide therapy using alpha Correspondence: I. Jambor, Department of Diagnostic Radiology, University of Turku, Kiinamyllynkatu 4-8, PO Box 52, Turku, Finland. Tel: Fax: ivjamb@utu.fi (Received 25 January 2015; accepted 5 March 2015) ISSN X print/issn X online 2015 Informa Healthcare DOI: / X

3 60 I. Jambor et al. emitting radium [4]. The degree of bone metastatic spread has been identified as an independent prognostic factor in PCa patients [5]. 99m Tc-methyl diphosphonate or 99m Tc-hydroxymethane diphosphonate ( 99m Tc-HDP) bone scintigraphy (BS) is the most widely used imaging modality for the detection of bone metastases. Unfortunately, several studies have demonstrated rather limited sensitivity (46 70%) and specificity (32 57%) of BS for the detection of bone metastases in BCa and PCa patients [6,7]. The use of single photon emission computed tomography (SPECT) alone or combined with computed tomography (CT) has shown to improve both sensitivity and specificity for the detection of bone metastases [6,8]. Positron emission tomography (PET) plays an important role in cancer imaging. Several studies indicate that 18 F-NaF PET/CT outperforms BS and SPECT in the detection of bone metastases [6,9 11]. Anatomical magnetic resonance imaging (MRI) has been used for the detection of bone metastases for more than 10 years [12 14]. Recent development in MR hardware and sequence optimization has resulted in whole body MRI (wbmri) capable of surveying the entire skeleton without irradiation over imaging times of min. Anatomical MRI with or without diffusion weighted imaging (DWI) is more sensitive and specific than BS for the detection of bone metastases [15 17]. In clinical practice anatomical MRI, with or without DWI, is commonly used to evaluate equivocal findings of nuclear medicine techniques, mainly BS. The cost of MRI is thus added to that of the nuclear medicine techniques and this approach may cause delay in the therapeutic decision process which delays treatment initiation such as the start of cancer therapy, further prolonging patients suffering. The aim of the current study was to compare the diagnostic accuracy of 99m Tc-HDP BS, 99m Tc-HDP SPECT, 99m Tc-HDP SPECT/CT, 18 F-NaF PET/ CT and whole body 1.5 Tesla (T) MRI, including DWI (wbmri DWI) for the detection of bone metastases in high risk BCa and PCa patients. Our goal was to reduce redundant bone imaging by studying the added value of tomographic and hybrid imaging techniques and wbmri DWI to standard approach which has been BS for many decades. Material and methods Patients Between February 2011 and March 2013, 26 BCa and 27 PCa patients were prospectively enrolled. The clinical trial registration number is NCT The study was approved by the local ethics committee and each patient gave written inform consent. Patients with high risk of bone metastases were enrolled if they met at least one of the following criteria: 1) high clinical suspicion of bone metastases: localized pain over bone area; 2) laboratory findings: elevated alkaline phosphates ( 105 U/l), elevated PSA or high PSA doubling time after prostatectomy, radiotherapy; 3) histopathologic findings: stage N3a or higher in BCa patients, stage T3a or higher and/ or Gleason score of 4 3 or higher in PCa patients. All BCa and PCa patients were referred for the evaluation of possible bone metastases after mastectomy/ prostatectomy and/or external radiotherapy. Fifteen (58%, 15/26) BCa and 13 (48%, 13/27) PCa patients had hormonal therapy at the time of imaging. All enrolled patients underwent 99m Tc-HDP BS, 99m Tc- HDP SPECT, 99m Tc-HDP SPECT/CT, 18 F NaF PET/CT and wbmri DWI within two weeks. Bone scintigraphy, SPECT and SPECT/CT Bone scan, SPECT and SPECT/CT examinations were performed using Symbia T6, True Point SPECT/ CT scanner (Siemens, Erlangen, Germany). The anterior and posterior views were collected 3 h after injection of 670 MBq 99m Tc-HDP using low-energy high-resolution (LEHR) collimators, scan speed 13 cm/min and matrix size Single-photon emission computed tomography combined with CT was acquired immediately after BS with the following parameters: three bed positions, LEHR collimators, 90 views, 9 s scanning time per view, matrix , zoom 1.0, energy window 140 kev 15% with lower scatter window, reconstruction using Flash 3D (Siemens, Erlangen, Germany) with five iterations and 10 subsets and CT with an effective mas 10 and 130 kvp. SPECT and SPECT/CT were performed from the tip of the head to the midthighs. 18 F-NaF PET/CT The patients received intravenous injection of (mean SD) MBq of 18 F-NaF diluted in 3 5 ml of saline as a 60 s bolus which was promptly flushed with saline. The PET data were acquired 64 6 (mean SD) min after the tracer injection using Discovery VCT PET/CT or PET/CT 690 scanner (General Electric Medical Systems, Milwaukee, WI, USA). A static emission scan was acquired over whole body with 3 min (PET/CT VCT scanner) and 2 min (PET/CT 690 scanner) acquisition time per bed position. The CT scan had a noise index of 25. The sinogram data were corrected for deadtime, decay, photon attenuation and reconstructed in a matrix. Image reconstruction followed a fully three-dimensional (3D) maximum likelihood ordered subsets expectation maximization algorithm

4 Detection of bone metastases in breast and prostate cancer patients 61 incorporating CT attenuation, random and scatter corrections with two iterations and 28 subsets. Whole body MRI Coronal T1-weighed (T1wi), coronal short tau inversion recovery (STIR) images and DWI data were acquired using a 1.5 T MR system (Avanto, Siemens, Erlangen, Germany) and surface array coils. The coronal images were collected from six table positions (table movement 250 mm) covering whole body with the following imaging parameters, T1wi: repetition time/echo time (TR/TE) 709/10 ms, field of view (FOV) 500 mm, slice thickness 7.0 mm, voxel size mm 3 ; STIR: TR/TE 14790/97 ms, inversion time 130 ms, FOV 500 mm, slice thickness 7.0 mm, voxel size mm 3. The axial DWI was performed using single shot spin echo sequence with 10 table positions (table movement 100 mm), TR/TE 6300/80 ms, SPAIR fat suppression, FOV 500 mm, slice thickness 5.0 mm, voxel size mm 3, three optimized diffusion directions (three-scan trace option on), b-values of 0, 150, 1000 s/mm 2.The lower limbs were covered in coronal orientation with 2 3 table positions (table movement 250 mm) and similar imaging parameters as axial DWI except of voxel size mm 3.The overall imaging time was approximately min. Data analysis and interpretation Each imaging modality was interpreted without the knowledge of other imaging results by three highly experienced nuclear medicine physicians and two radiologists. The reviewers were only aware of the fact that the patients are at high risk of bone metastases. Lesions were graded as highly suspicious for being metastases, equivocal or benign. On BS and SPECT scans, lesions were categorized as benign when they were located around joints, hot osteophytes, vertically involving several ribs (suggesting fracture), H-shaped pelvic abnormal, bursitis, avulsion injury, tendinitis [9,10]. The vertebral lesions were considered as highly suspicious when they involved posterior aspect, pedicle or the whole vertebral body. Anatomical CT images of PET/CT and SPECT/CT were used to categorize uptakes as benign or highly suspicious due to corresponding morphologic findings. Typical benign lesions according to CT data were bone cysts, degenerative lesions (e.g. around joints), and fractures. When the tracer uptake was located on osteoblastic, osteolytic or mixed lesion, the lesion was marked as highly suspicious based on PET/CT and/ or SPECT/CT. Lesions associated with a tracer uptake and not typical benign or malignant changes on CT were considered as equivocal. A lesion was considered as highly suspicious on MRI if a focal or diffusion low signal intensity (SI) was present on T1wi with the corresponding intermediate or high SI on STIR and/or restricted diffusion on DWI. Typical benign lesions and/or sclerosis were interpreted according to previously published criteria [14,18]. Trace images (mean of three diffusion directions) were evaluated visually in conjunction with anatomical T1wi and STIR. No quantitative cut-off values were used for DWI. The bone findings were compared on patient-, region- and lesion-level. In the region-based analysis, the skeleton was divided into five regions: head, thorax and ribs, spine, pelvis and limbs. In the lesion-based analysis, only lesions which were highly suspicious or equivocal on at least one imaging modality were included. In addition, maximum of five lesions with the highest agreement between modalities per anatomical location (five locations as defined in the region-based analysis) were included in the lesion-based analysis. Best valuable comparator (BCV) The findings of each imaging modality were compared with best valuable comparator (BVC) in order to define their nature [7,19]. Consensus reading of all imaging modalities and follow-up data of clinical, imaging and laboratory results were used to define BVC. The mean standard deviation follow-up time of 53 enrolled patients was 15 7 months while the range was 6 32 months. In total, 74 follow-up imaging examinations were performed consisting of 21 99m Tc-HDP BS, one 99m Tc-HDP SPECT/CT, 38 CT, four 18 F-NaF PET/CT, three 18 F-NaF PET/ MRI, and seven MRI examinations. Three BCa and two PCa patients died during the follow-up period. Imaging follow-up data were available for 14 BCa and 11 PCa patients. These studies were mainly performed in the patients with disconcordant findings among the different imaging modalities while no follow-up imaging was done in 10 BCa and 10 PCa patients with no highly suspicion lesions on any of the imaging modalities. These patients did not show any signs of progression during the follow-up period and were considered as true negative. Three PCa patients had no imaging follow-up examinations while having highly suspicious lesion(s) on BS and/ or SPECT which were considered as false positive based on the consensus imaging findings of the initial imaging examinations. Highly suspicious lesions were considered as true positive if the lesions on the consensus reading of the initial imaging examinations and/or follow-up examinations were positive for bone metastases. Lesions were considered as false positive if rated as highly suspicious on one modality

5 62 I. Jambor et al. Table I. Patient-based analysis. Optimistic analysis Pessimistic analysis Sensitivity Specificity Accuracy AUC Sensitivity Specificity Accuracy AUC BS 79 (p 0.001) 91 (p 0.16) (p 0.05) 85 (p 0.08) 59 (p 0.01) (p 0.01) SPECT 89 (p 0.16) 80 (p 0.01) (p 0.01) 95 (p 0.32) 56 (p 0.01) (p 0.01) SPECT/CT 89 (p 0.16) 94 (p 0.32) (p 0.12) 95 (p 0.32) 88 (NA) (p 0.12) PET/CT 95 (p 0.32) 97 (NA) (p 0.42) 100 (NA) 82 (p 0.32) (p 0.15) wbmri DWI 100 ( * ) 97 ( * ) ( * ) 100 ( * ) 88 ( * ) ( * ) Sensitivity, specificity and accuracy values are displayed in %. In each group, two-sided p-values are displayed with reference to the modality marked by ( * ). BS, 99m Tc-hydroxymethane diphosphonate bone scintigraphy; NA, not applicable; PET/CT, 18 F-NaF positron emission tomography combined with computed tomography; SPECT, 99m Tc-hydroxymethane diphosphonate single photon emission computed tomography; SPECT/CT, 99m Tc-hydroxymethane diphosphonate single photon emission computed tomography combined with computed tomography; wbmri DWI, whole body MRI (T1-weighted imaging, STIR) including diffusion weighted imaging. but as benign on consensus reading of the initial imaging examinations and/or no signs of active disease in any of the follow-up examinations was discovered. False negative finding was made when a reader found no lesion or marked as a benign lesion and the consensus reading of the initial imaging examinations and/or follow-up examinations showed bone metastases. If no imaging follow-up data were available (12 BCa and 16 PCa patients), patients with clinical, laboratory follow-up data suggesting progression of the metastatic disease, as well as consensus reading of the initial imaging examinations suggesting metastases, were considered to have metastatic disease at the time of initial imaging (two BCa and three PCa patients). Statistical analysis Equivocal findings of the imaging modalities were classified either as suggestive for metastases ( pessimistic analysis ) or suggestive for non-metastatic origin ( optimistic analysis ). Sensitivity and specificity values of patient-, region- and lesion-based analyses were compared using McNemar test [20] and two-sided p-values were calculated. In regionbased analysis, diagnostic accuracy values for the detection of bone metastases [sensitivity, specificity, accuracy and area under the curve (AUC)] were calculated from all ROIs which were pooled into one group. Moreover, receiver operation characteristic curves (ROC) analysis was performed using bootstraps [21] to account for within-patient correlations. AUC values were calculated using the trapezoid rule and compared using a method described by Hanley and McNeil [22], two-sided p-values were calculated. Bootstrap samples were constructed by stratifying patients based on overall cancer level (BCa/PCa present or not) and drawing patients as the independent units with replacement from these groups (BCa/PCa present or not) [21]. P values 0.05 were considered statistically significant. All statistical analyses were performed using in-house written Matlab codes (Mathworks Inc., Natick, MA, USA). The Matlab codes as well as all MRI sequences are freely available upon request. Results Patient-based analysis In patient-based analysis, 19 (36%, 19/53) patients, 11 (42%, 11/26) BCa and eight (30%, 8/27) PCa patients, had presence of bone metastases based on BVC. BS had significantly lower sensitive and AUC values than 99m Tc-HDP SPECT/CT, 18 F-NaF PET/ CT and wbmri DWI (Table I; Figures 1, 2). When equivocal lesions were considered as suggestive for bone metastases (pessimistic analysis), 99m Tc-HDP BS and 99m Tc-HDP SPECT had significantly lower specificity values than 99m Tc-HDP SPECT/CT, 18 F-NaF PET/CT and wbmri DWI. Differences in sensitivity, specificity, AUC values of 99m Tc-HDP Figure 1. 99m Tc-hydroxymethane diphosphonate bone scintigraphy, anterior (A) and posterior view (B), and SPECT (C) of 69-yearold prostate cancer patient with serum PSA values of 97 ng/ml and Gleason score of 5 4 based on systematic biopsy. One equivocal lesions (10th rib, marked by black arrows) was present on 99m Tc-hydroxymethane diphosphonate bone scintigraphy while focal uptake was present on SPECT in the 10th rib on the left side and right lower part of sacral bone (marked by black arrows) suggesting bone metastases.

6 Detection of bone metastases in breast and prostate cancer patients 63 analyses (Table II). In pessimistic analysis, 18 F-NaF PET/CT and wbmri DWI were significantly more sensitive than 99m Tc-HDP SPECT/CT while differences in specificity and AUC values did not reach the level of statistically significance. In contrast, when equivocal lesions were considered as not suggestive for bone metastases (optimistic analysis), the differences in sensitivity, specificity and AUC values of 99m Tc-HDP SPECT/CT, 18 F-NaF PET/CT and wbmri DWI did not reach statistical significance. Figure 2. Coronal 18 F-NaF PET and DWI, inverted signal intensity of trace image b-value of 1000 s/mm 2, of the same 69- year-old prostate cancer patient as displayed in Figure 1. The lesions in the 10th rib on the left side and right lower part of sacral bone (marked by black arrows) were highly suspicious for being bone metastases similarly to SPECT (Figure 1) and SPEC/CT (not shown) while the lesion in sacral bone was missed and the in the 10th rib was considered as equivocal on bone scintigraphy. These highly suspicious lesions were suggestive to be bone metastases based on clinical and imaging follow-up, demonstrating a limitation of bone scintigraphy for the detection of bone metastases in prostate cancer patients. SPECT/CT, 18 F-NaF PET/CT and wbmri DWI did not reach statistical significance. Region-based analysis In region-based analysis, 58 (22%, 58/265) ROIs, 39 (30%, 39/130) BCa and 19 (14%, 19/135) PCa ROIs, had presence of bone metastases based on BVC. 99m Tc-HDP BS and 99m Tc-HDP SPECT had significantly lower sensitive and AUC values than 99m Tc-HDP SPECT/CT, 18 F-NaF PET/CT and wbmri DWI in both pessimistic and optimistic Lesion-based analysis In total, 234 lesions, 172 in BCa and 62 in PCa patients, were highly suspicious or equivocal in at least one imaging modality and no more than five lesions per region with the highest agreement between modalities, as defined in the region-based analysis, were included. Of these 234 lesions, 159 (68%, 159/234) lesions, 123 (72%, 123/172) in BCa and 36 (58%, 36/62) in PCa patients, were considered to be metastatic bone lesions based on BVC. 18 F-NaF PET/CT and wbmri DWI were significantly more sensitive in the lesion-based analysis than 99m Tc-HDP BS, 99m Tc- HDP SPECT, 99m Tc-HDP SPECT/CT (Table 3). Moreover, 18 F-NaF PET/CT and wbmri DWI had significantly higher specificity values than 99m Tc-HDP BS and 99m Tc-HDP SPECT. 18 F-NaF PET/CT and wbmri DWI had similar sensitivity, specificity, accuracy and AUC values (Figures 3 5). The number of equivocal lesions was 50, 44, 5, 6 and 4 in 99m Tc-HDP BS, 99m Tc-HDP SPECT, 99m Tc-HDP SPECT/CT, 18 F-NaF PET/CT and wbmri DWI readings, respectively. These lesions were present in 22 (42%, 22/53), 20 (38%, 20/53), 5 (9%, 5/53), 4 (11%, 4/53), and 3 (8%, 3/53) patients of 99m Tc-HDP BS, 99m Tc-HDP SPECT, SPECT/CT, 18 F-NaF PET/CT, and wbmri DWI readings, respectively. Furthermore, in 13 (25%, 13/53), 9 (17%, 9/53), 4 (8%, 4/53), 4 (8%, 4/53) and 2 (4%, 2/53) patients only equivocal lesions with Table II. Region-based analysis. Optimistic analysis Pessimistic analysis Sensitivity Specificity Accuracy AUC Sensitivity Specificity Accuracy AUC BS 62 (p 0.001) 98 (p 0.32) (p 0.001) 72 (p 0.01) 89 (p 0.01) (p 0.001) SPECT 74 (p 0.01) 94 (p 0.01) (p 0.01) 86 (p 0.05) 87 (p 0.001) (p 0.01) SPECT/CT 85 (p 0.05) 99 (p 0.32) (p 0.24) 85 (p 0.01) 98 (NA) (p 0.07) PET/CT 93 ( * ) 99 ( * ) ( * ) 97 ( * ) 97 (p 0.52) ( * ) wbmri DWI 91 (p 0.65) 99 (p 0.32) (p 0.92) 91 (p 0.25) 98 ( * ) (p 0.36) Sensitivity, specificity and accuracy values are displayed in %. In each group, two-sided p-values are displayed with reference to the modality marked by ( * ). BS, 99m Tc-hydroxymethane diphosphonate bone scintigraphy; NA, not applicable; PET/CT, 18 F-NaF positron emission tomography combined with computed tomography; SPECT, 99m Tc-hydroxymethane diphosphonate single photon emission computed tomography; SPECT/CT, 99m Tc-hydroxymethane diphosphonate single photon emission computed tomography combined with computed tomography; wbmri DWI, whole body MRI (T1-weighted imaging, STIR) including diffusion weighted imaging.

7 64 I. Jambor et al. Figure 3. 99m Tc-hydroxymethane diphosphonate bone scintigraphy, anterior (A) and posterior view (B), and SPECT (C) of 62-yearold breast cancer patient with elevated alkaline phosphatase and localized pain in upper back. Lesions in the 5th and 7th rib (marked by black arrows) on the left side are highly suspicions of being bone metastases based on bone scintigraphy and SPECT. The lesions (marked by half filled arrows) in the fifth lumbal vertebra and 9th rib on the left side were considered as equivocal on prospective SPECT reading while the increased uptake (marked by black arrowhead) in the left side of thoracic spine at the level of X was considered as a benign uptake on prospective SPECT reading. or without benign lesions were present in 99m Tc- HDP BS, 99m Tc-HDP SPECT, 99m Tc-HDP SPECT/ CT, 18 F-NaF PET/CT, wbmri DWI readings, respectively. Change in patient management In two PCa patients, the detection of bone metastases discovered only by SPECT/CT, 18 F-NaF PET/ CT and wbmri DWI resulted in the change of treatment (start of chemotherapy). In one BCa patient, hormonal treatment was initiated due to detection of bone metastases detected by 18 F-NaF PET/CT and wbmri DWI. Due to detection of liver and mesenteric metastases only by wbmri DWI, treatment plan was changed in one BCa and one PCa patients, respectively. These lesions were confirmed to be metastases by the imaging and clinical follow-up. Figure 4. Coronal 99m Tc-hydroxymethane diphosphonate SPECT/ CT (A) 18 F-NaF PET/CT (B) and T1-weighted image (C) of the same 62-year-old breast cancer patient as displayed in Figure 3. In addition to the lesions in the 5th and 7th rib on the left side detected on 99m Tc-hydroxymethane diphosphonate bone scintigraphy and SPECT, lytic lesion was present in the body of the 5th lumbal vertebra (marked by white arrow). Figure 5. Maximum intensity projection of 18 F-NaF PET (A) and DWI (B), inverted signal intensity of trace image b-value of 1000 s/mm 2, of the same 62-year-old breast cancer patient as displayed in Figures 3 and 4. In addition to the lesions in the 5th, 7th rib on the left side and L5 (marked by black arrows) detected by SPECT/CT, two additional highly suspicion lesions (marked by black arrows) in the 6th and 9th rib on the right side were detected by 18 F-NaF PET/CT and whole body MRI, including DWI. All of the highly suspicious lesions were suggestive to be bone metastases based on clinical and imaging follow-up, showing limited accuracy of bone scintigraphy for the detection of bone metastases in breast cancer patients. Magnetic susceptibility related artifacts are present on DWI in the right hip due to presence of metal hip prosthesis. Estimates of price range for each of the imaging modalities, patient examination time, and reading time are shown in Table IV. Please note that the price ranges are just estimates based on data available from Finland s healthcare system. Nevertheless, it can be seen that 18 F-NaF PET/CT is associated with the higher cost which is about four times the cost of 99m Tc-HDP SPECT/CT and wbmri DWI. Moreover, the reading times (time a reader spends reporting the exam) are comparable between 99m Tc-HDP SPECT/CT, 18 F-NaF PET/CT, and wbmri DWI. Discussion Accurate detection of bone metastases in BCa and PCa patients is vital for treatment planning and patient prognosis. In the current study, we have evaluated diagnostic accuracy of 99m Tc-HDP BS, SPECT, SPECT/CT, 18 F-NaF PET/CT, and wbmri, including DWI, for the detection of bone metastases in high risk BCa and PCa patients. 99m Tc-HDP SPECT/CT, 18 F-NaF PET/CT, and wbmri, including DWI, had significantly higher AUC values for the detection of bone metastases than 99m Tc-HDP BS and SPECT. Furthermore, 18 F-NaF PET/CT and wbmri DWI demonstrated significantly higher sensitivity and AUC values in lesion-based analysis than conventional nuclear medicine techniques (Figures 3, 4 and 5). To our knowledge, this is the

8 Detection of bone metastases in breast and prostate cancer patients 65 Table III. Lesion-based analysis. Optimistic analysis Pessimistic analysis Sensitivity Specificity Accuracy AUC Sensitivity Specificity Accuracy AUC BS 54 (p 0.001) 88 (p 0.01) (p 0.001) 69 (p 0.001) 72 (p 0.001) (p 0.001) SPECT 71 (p 0.001) 79 (p 0.01) (p 0.001) 77 (p 0.001) 52 (p 0.001) (p 0.001) SPECT/CT 81 (p 0.001) 96 (NA) (p 0.04) 82 (p 0.001) 94 (NA) (p 0.02) PET/CT 94 (p 0.78) 96 ( * ) ( * ) 96 ( * ) 89 (p 0.20) (p 0.38) wbmri DWI 95 ( * ) 95 (p 0.65) (p 0.91) 95 (p 0.76) 94 ( * ) ( * ) Sensitivity, specificity and accuracy values are displayed in %. In each group, two-sided p-values are displayed with reference to the modality marked by ( * ). BS, 99m Tc-hydroxymethane diphosphonate bone scintigraphy; NA, not applicable; PET/CT, 18 F-NaF positron emission tomography combined with computed tomography; SPECT, 99m Tc-hydroxymethane diphosphonate single photon emission computed tomography; SPECT/CT, 99m Tc-hydroxymethane diphosphonate single photon emission computed tomography combined with computed tomography; wbmri DWI, whole body MRI (T1-weighted imaging, STIR) including diffusion weighted imaging. first prospective clinical trial directly comparing the diagnostic accuracy of 99m Tc-HDP BS, SPECT, SPECT/CT, 18 F-NaF PET/CT, and wbmri, including DWI, for the detection of bone metastases in high risk BCa and PCa patients. The current study enrolled high risk BCa and PCa with the aim to compare diagnostic accuracy for the detection of bone metastases rather than to evaluate the impact of different imaging modalities on diagnosis and treatment planning. Similarly to recent study by Lecouvet et al. [19], our end points were pure diagnostic with special focus on bone metastases. Whole body MRI, including DWI, demonstrated high diagnostic performance for the detection of bone metastases and was significantly more sensitive than BS, which is still the most widely used imaging modality for the detection of bone metastases. Accurate evaluation of bone metastatic spread is becoming more important with introduction of novel therapeutic option [4]. In the current study, BS and SPECT missed four and two patients with bone metastases of 53 in per patient analysis, respec- Table IV. Estimated price ranges, examination times, and reading times of the imaging modalities. Estimated price ranges of the imaging modalities (EUR) Examination time (min) Reading time (min) BS SPECT SPECT/CT PET/CT wbmri DWI BS, 99m Tc-hydroxymethane diphosphonate bone scintigraphy; PET/CT, 18 F-NaF positron emission tomography combined with computed tomography; SPECT, 99m Tc-hydroxymethane diphosphonate single photon emission computed tomography; SPECT/CT, 99m Tc-hydroxymethane diphosphonate single photon emission computed tomography combined with computed tomography; wbmri DWI, whole body MRI (T1-weighted imaging, STIR) including diffusion weighted imaging. tively. However, only two of those four patients had change in the treatment management (initiation of hormonal therapy) since the remaining two patients had hormonal therapy at the time of imaging. This could potentially raise a question, what is the current clinical additive value of scanning high risk BCa and PCa patients with more advanced methods? The major difference, however, is in the diagnostic certainty or uncertainty with BS and SPECT due to high numbers of equivocal lesions compared with SPECT/CT, PET/CT or wbmri DWI. Equivocal lesions of BS were present in 39% (22/53) of patients and in 25% (13/53) of patients only equivocal lesions with or without benign lesions were present. Similarly, 38% (20/53) of patients had equivocal lesions on SPECT and 17% (9/53) of patients had only equivocal lesions with or without benign lesions on SPECT. This group of patients causes a major diagnostic dilemma and therefore often require further imaging studies to reveal the nature of lesions. This leads to unnecessary delays in diagnosis causing patient discomfort and extra costs. Repeated imaging studies aiming to detect bone metastases are commonly needed in treatment planning of patients and cumulative radiation dose could become an issue, at least among the youngest patients. Our finding of wbmri DWI and 18 F-NaF PET/ CT being superior to conventional nuclear imaging techniques for discovering bone metastases is in line with previous studies [6,7,9,19]. Considering the cost (Table IV), availability and radiation dose, wbmri DWI may be a preferred choice in comparison with 18 F-NaF PET/CT. Whole body MRI, including DWI, was as accurate as 18 F-NaF PET/CT for the detection of bone metastases in the current study. Moreover, wbmri DWI can potentially provide useful information concerning soft tissues [19]. Our findings support the use of wbmri DWI as a single-step imaging method for the detection of both soft and bone metastases in high risk BCa and PCa patients. The strategy of single-step detection

9 66 I. Jambor et al. of metastases using wbmri DWI in high risk PCa was previously suggested by Lecouvet et al. [19]. The wide use of this imaging modality for the detection of bone metastases in high risk BCa and PCa patient could still be partly limited by rather long imaging time of min. However, further development of MR hardware and sequences could allow performing robust wbmri DWI in less than 30 min. In the current study, DWI was evaluated in conjunction with T1wi and STIR to reduce number of false positive lesions [15] and only visual approach was used for DWI data, trace images. Quantitative evaluation of DWI could potentially allow monitoring of therapy response and early detection of responders from those patients who need change in treatment [23], potentially strengthening a role of wbmri DWI as single-step imaging modality for detection of bone metastases in high risk BCa and PCa patients. Correct assessment of true nature of the lesions is the main limitation of the current study because histological confirmation was not available. In addition to consensus reading of all imaging modalities, clinical, and imaging follow-up of at least six months were used to define true nature of the lesions detected by each of the modalities. This approach is similar to many other studies focusing on the detection of bone and/or lymph node metastases [7,19,24,25]. However, this can result in overestimation of diagnostic accuracy of the most accurate imaging modality [19]. Differences in experience of 99m Tc-HDP BS, 99m Tc- HDP SPECT, 99m Tc-HDP SPECT/CT, 18 F-NaF PET/CT and wbmri DWI readers could potentially also have influence on the results. However, all of the readers are highly experienced nuclear medicine physicians and/or radiologists with at least five years of experiences in the detection of bone metastases using that particular imaging modality. The current study is further limited by relatively small number of patients. In the current study, 99m Tc-HDP SPECT/CT, 18 F-NaF PET/CT and wbmri DWI were significantly more sensitive than 99m Tc-HDP BS, 99m Tc- HDP SPECT for the detection of bone metastases in high risk BCa and PCa patients. Whole body MRI, including DWI, was as accurate as 18 F-NaF PET/CT for the detection of bone metastases in high risk BCa and PCa patients. In the context of nuclear medicine techniques, 99m Tc-HDP SPECT/CT was superior to 99m Tc-HDP BS, and SPECT, especially SPECT/CT having less equivocal findings. Acknowledgments This study was financially supported by grants from the Instrumentarium Research Foundation, Sigrid Jusélius Foundation, Turku University Hospital, TYKS-SAPA research fund, Finnish Cancer Society, and Finnish Cultural Foundation. Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper. References [1] Siegel R, Naishadham D, Jemal A. Cancer statistics, CA Cancer J Clin 2012;62: [2] Roodman GD. Mechanisms of bone metastasis. N Engl J Med 2004;350: [3] Zerbib M, Zelefsky MJ, Higano CS, Carroll PR. Conventional treatments of localized prostate cancer. Urology 2008;72:S [4] Parker C, Nilsson S, Heinrich D, Helle SI, O Sullivan JM, Fossa SD, et al. Alpha emitter radium-223 and survival in metastatic prostate cancer. N Engl J Med 2013;369: [5] Rigaud J, Tiguert R, Le NL, Karam G, Glemain P, Buzelin JM, et al. Prognostic value of bone scan in patients with metastatic prostate cancer treated initially with androgen deprivation therapy. J Urol 2002;168: [6] Even-Sapir E, Metser U, Mishani E, Lievshitz G, Lerman H, Leibovitch I. The detection of bone metastases in patients with high-risk prostate cancer: 99mTc-MDP Planar bone scintigraphy, single- and multi-field-of-view SPECT, 18Ffluoride PET, and 18F-fluoride PET/CT. J Nucl Med 2006;47: [7] Lecouvet FE, Geukens D, Stainier A, Jamar F, Jamart J, d Othee BJ, et al. Magnetic resonance imaging of the axial skeleton for detecting bone metastases in patients with high-risk prostate cancer: Diagnostic and cost-effectiveness and comparison with current detection strategies. J Clin Oncol 2007;25: [8] Schirrmeister H, Glatting G, Hetzel J, Nussle K, Arslandemir C, Buck AK, et al. Prospective evaluation of the clinical value of planar bone scans, SPECT, and (18) F-labeled NaF PET in newly diagnosed lung cancer. J Nucl Med 2001;42: [9] Even-Sapir E, Metser U, Flusser G, Zuriel L, Kollender Y, Lerman H, et al. Assessment of malignant skeletal disease: Initial experience with 18F-fluoride PET/CT and comparison between 18F-fluoride PET and 18F-fluoride PET/CT. J Nucl Med 2004;45: [10] Even-Sapir E. Imaging of malignant bone involvement by morphologic, scintigraphic, and hybrid modalities1. J Nucl Med 2005;46: [11] Iagaru A, Mittra E, Dick DW, Gambhir SS. Prospective evaluation of (99m)Tc MDP scintigraphy, (18)F NaF PET/CT, and (18)F FDG PET/CT for detection of skeletal metastases. Mol Imaging Biol 2012;14: [12] Walker R, Kessar P, Blanchard R, Dimasi M, Harper K, DeCarvalho V, et al. Turbo STIR magnetic resonance imaging as a whole-body screening tool for metastases in patients with breast carcinoma: Preliminary clinical experience. J Magn Reson Imaging 2000;11: [13] Steinborn MM, Heuck AF, Tiling R, Bruegel M, Gauger L, Reiser MF. Whole-body bone marrow MRI in patients with metastatic disease to the skeletal system. J Comput Assist Tomogr 1999;23: [14] Vanel D, Bittoun J, Tardivon A. MRI of bone metastases. Eur Radiol 1998;8:

10 Detection of bone metastases in breast and prostate cancer patients 67 [15] Grankvist J, Fisker R, Iyer V, Frund ET, Simonsen C, Christensen T, et al. MRI and PET/CT of patients with bone metastases from breast carcinoma. Eur J Radiol 2012;81:e13 8. [16] Goudarzi B, Kishimoto R, Komatsu S, Ishikawa H, Yoshikawa K, Kandatsu S, et al. Detection of bone metastases using diffusion weighted magnetic resonance imaging: Comparison with (11)C methionine PET and bone scintigraphy. Magn Reson Imaging 2010;28: [17] Ghanem N, Uhl M, Brink I, Schafer O, Kelly T, Moser E, et al. Diagnostic value of MRI in comparison to scintigraphy, PET, MS-CT and PET/CT for the detection of metastases of bone. Eur J Radiol 2005;55: [18] Schmidt GP, Schoenberg SO, Schmid R, Stahl R, Tiling R, Becker CR, et al. Screening for bone metastases: Whole-body MRI using a 32-channel system versus dual-modality PET-CT. Eur Radiol 2007;17: [19] Lecouvet FE, El MJ, Collette L, Coche E, Danse E, Jamar F, et al. Can whole-body magnetic resonance imaging with diffusion-weighted imaging replace Tc 99m bone scanning and computed tomography for single-step detection of metastases in patients with high-risk prostate cancer? Eur Urol 2012;62: [20] Trajman A, Luiz RR. McNemar chi2 test revisited: Comparing sensitivity and specificity of diagnostic examinations. Scand J Clin Lab Invest 2008;68: [21] Rutter CM. Bootstrap estimation of diagnostic accuracy with patient-clustered data. Acad Radiol 2000;7: [22] Hanley JA, McNeil BJ. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 1983;148: [23] Messiou C, Collins DJ, Giles S, de Bono JS, Bianchini D, de Souza NM. Assessing response in bone metastases in prostate cancer with diffusion weighted MRI. Eur Radiol 2011;21: [24] Lecouvet FE, Vande Berg BC, Malghem J, Omoumi P, Simoni P. Diffusion-weighted MR imaging: Adjunct or alternative to T1-weighted MR imaging for prostate carcinoma bone metastases? Radiology 2009;252:624. [25] Venkitaraman R, Cook GJ, Dearnaley DP, Parker CC, Khoo V, Eeles R, et al. Whole-body magnetic resonance imaging in the detection of skeletal metastases in patients with prostate cancer. J Med Imaging Radiat Oncol 2009;53:241 7.