Apparent Diffusion Coefficient Maps Integrated in Whole-Body MRI Examination for the Evaluation of Tumor Response to Chemotherapy in Patients with Multiple Myeloma Pietro Andrea Bonaffini, MD, Davide Ippolito, MD, Alessandra Casiraghi, MD, Valeria Besostri, MD, Cammillo Talei Franzesi, MD, Sandro Sironi, MD Rationale and objectives: To determine the diagnostic value of apparent diffusion coefficient (ADC) maps in the assessment of response to chemotherapy in patients with multiple myeloma (MM). Materials and methods: Fourteen patients (seven women) with MM underwent whole-body magnetic resonance imaging (WB-MRI) study on a 1.5T scanner, before and after chemotherapy. DWI with background body signal suppression (DWIBS) sequences (b values: 0, 500, and 1000 mm2/sec) were qualitatively analyzed, along with T1 turbo spine echo and short tau inversion recovery T2-weighted images, to evaluate bone lesions. On ADC maps, regions of interest were manually drawn along contours of lesions. The ADC values percentage variation (DADC) before (MR1) and after (MR2) chemotherapy were calculated and compared between responders (11 of 14) and nonresponders (3 of 14). The percentage of plasma cells by the means of the bone marrow aspirate was evaluated as parameter for response to chemotherapy. Results: Twenty-four lesions, hyperintense on DWIBS as compared to normal bone marrow, were evaluated. In responder group, the mean ADC values were 0.63 0.24 10-3 mm2/s on MR1 and 1.04 0.46 10-3 mm2/s on MR2; partial or complete signal intensity decrease during follow-up on DWIBS was found along with a reduction of plasma cells infiltration in the bone marrow. The mean ADC values for nonresponders were 0.61 0.05 10-3 mm2/s on MR1 and 0.69 0.09 10-3 mm2/s on MR2. The mean variation of DADC in responders (D = 66%) was significantly different (P < .05) than in nonresponders (D = 15%). Conclusions: WB-MRI with DWIBS sequences, by evaluating posttreatment changes of ADC values, might represent a complementary diagnostic tool in the assessment of response to chemotherapy in MM patients. Key Words: Multiple myeloma; chemotherapy; ADC; whole-body MRI. ªAUR, 2015
I
n the past decades, magnetic resonance imaging (MRI) has shown a high diagnostic and prognostic value in monoclonal plasma cell disorders (1–4). It has also been included as additional imaging tool for diagnosis and staging protocols of multiple myeloma (MM) (2,4), but, to date,
Acad Radiol 2015; 22:1163–1171 From the School of Medicine, University of Milano-Bicocca, Via Pergolesi 33, 20900 Monza, Italy (P.A.B., D.I., A.C., V.B., C.T.F., S.S.); Department of Diagnostic Radiology, H. San Gerardo, Monza, Italy (P.A.B., D.I., A.C., V.B., C.T.F., S.S.); and Department of Health Sciences, Tecnomed Foundation, University of Milano-Bicocca, Monza, Italy (P.A.B.). Received November 23, 2014; accepted May 27, 2015. Address correspondence to: P.A.B. e-mail: [email protected] ªAUR, 2015 http://dx.doi.org/10.1016/j.acra.2015.05.011
there is insufficient evidence to recommend routine MRI after treatment because most of the findings are not specific (1,2,5). In MM follow-up, the role of imaging is still limited, relying on skeletal survey although bone marrow aspirate and laboratory parameters (serum and urinary M-protein measurements) are the mainstay of the evaluation of response to chemotherapy. According to the most recent guidelines, plain films should be repeated after clinical or laboratory evidence of disease progression (2,5–7), but on conventional radiographs, lytic bone lesions rarely demonstrate signs of healing. Moreover, the evidence of new vertebral compressions may not indicate disease progression, occurring even after an effective therapy (1,5). Therefore, an accurate imaging technique able to detect treatment-related changes 1163
BONAFFINI ET AL
could increase the ability to monitor MM patients, even those with reduced or absent M-protein secretion. Diffusion-weighted imaging (DWI) is an additional and noninvasive tool that allows visualization and measurement of microscopic Brownian motion (diffusivity) of water molecules in the tissues. Thus, MRI coupled with DWI can provide within a single examination, morphological details along with functional information, related to tissue cellularity, to the cell membrane integrity and to intracellular/extracellular space–volume ratio (8–11). The introduction of DWI with background body signal suppression (DWIBS) technique (12) allows the acquisition, during free breathing, of volumetric and thin-sliced diffusion-weighted images of the whole body (8), making it applicable even in MM patients. Along with visual and qualitative analysis, diffusion can be also quantified in biological tissues by means of the so-called apparent diffusion coefficient (ADC) (8,9). Several studies have shown the ability of DWI in the detection and characterization of different pathologic processes (acute ischemic stroke, inflammatory lesions, malignant tumors) and in the differentiation of postoperative changes from residual active tumor or recurrence (11,13–16). DWI role in the evaluation of response to treatment has to be still clearly defined, and to knowledge, there are few studies that evaluated it in MM (7,17,18). Nevertheless, usefulness of DWI in this issue is expected because of high cellular load replacing hematopoietic bone marrow that characterized MM (7,19,20). The purpose of this study was to assess the value of quantitative analysis of ADC maps, obtained by DWIBS sequences, in the evaluation of response to chemotherapy in MM patients. MATERIALS AND METHODS Study Population
Institutional ethical committee approval was obtained for this retrospective analysis of a total of 28 MRI studies performed in 14 patients (seven women; average age, 60 years; age range, 40–76 years). The main inclusion criteria were patients referred to our department with histologic diagnosis of MM by means of bone marrow aspirate, undergoing whole-body MRI (WB-MRI) examination, and who were eligible to treatment. Patients with other malignant tumors (ie, lung cancer) and who were not able to properly complete the whole MRI study were excluded. Every patient underwent a WB-MRI study both for initial staging (MR1) and on the average 7 months (range, 2–17) after chemotherapy treatment (MR2). Informed consent for MRI was obtained from all the patients. All the included patients were treated with standard protocols, which included chemotherapy drugs (bortezomib, thalidomide, vincristine, doxorubicin, melphalan), alone or combined with the administration of dexamethasone. 1164
Academic Radiology, Vol 22, No 9, September 2015
MRI Technique
Every WB-MRI examination was performed on a 1.5T scanner (Achieva; Philips Medical Systems, Best, The Netherlands), using a built-in body coil combined with stepping table technique, as supplied by the manufacturer and during free but regular breathing. No patient preparation was required, and no contrast material was injected. The patients were positioned in the magnet supine and head first. After the acquisition of survey images of the whole body (range, 5 to 7, depending on patient’s height), the WB-MRI protocol consisted of the following sequences (Table 1): T2 short tau inversion recovery (STIR; turbo factor, 30; repetition time [TR], 1400; echo time [TE], 64; inversion time [TI], 165; flip angle [FA], 90 ; number of signals acquired [NSA], 2) and T1 turbo spine echo (TSE; TR, 1000; TE, 17.50; FA, 90 ; NSA, 1) sequences, acquired on the coronal plane from the skull vertex to the feet (T1: voxel size 1.27 1.82 mm3; STIR: voxel size 1.58 1.97 mm3). The field of view (FOV) was manually adapted to the major circumference of the patient. The same sequences were also acquired on the sagittal plane covering the whole axial skeleton (T1: voxel size 1.17 1.53 mm3; STIR: voxel size 1.38 1.75 mm3). On the same coverage area, DWIBS sequences were acquired on the axial plane to avoid distortion artifacts, using three different b values (b = 0, b = 500 e b = 1000 s/mm2) and with the following parameters: echo-planar imaging (EPI), TR, 1400; TE, 66; FA, 90 ; NSA, 2; voxel size 5 4.97 mm3. To obtain fat suppression, DWIBS was combined with an STIR prepulse. All the study sequences were acquired during free breathing, with slice thickness of 4and 1-mm gap. The total acquisition time was about 60–75 minutes, depending on patient’s height. At the end of the study, every imaged district was merged using software integrated in the scanner, generating coronal whole-body T1, T2 STIR, and DWIBS reconstructions (Fig. 1). Moreover, inverting the gray scale of DWIBS images on reporting workstation, ‘‘PET-like’’ images were available for radiologists at the time of evaluation of examination. Patients and Imaging Analysis
At the time of diagnosis and during follow-up, the percentage of plasma cells by the means of the bone marrow aspirate was evaluated, and the response to treatment was established as a reduction of bone marrow infiltration (6). When histologic analysis was not available at our Institution’s electronic database (2 of 14 patients, 1 responder), the response to treatment was established according to the clinical evaluation by medical records. At the first examination, axial DWIBS sequences were compared with standard T1 and T2 STIR images in five different anatomic districts (skull, spine, sternum and ribs, pelvis, upper and lower extremities) to detect the primary
Academic Radiology, Vol 22, No 9, September 2015
ADC MAPS FOR RESPONSE TO CHEMOTHERAPY IN MM
TABLE 1. Whole-Body Magnetic Resonance Imaging Protocol (1.5T), Acquired with Patient Positioned Supine and Using Built-In Body Coil Combined with Stepping Table Technique Sequence Slice thickness, mm Plane of acquisition Range of scan b values TR TE TI FA NSA Voxel size
T2 STIR
T1 TSE
DWIBS
4 Coronal (C) and sagittal (S) Coronal: skull vertex to feet. Sagittal: axial skeleton / 1400 64 165 90 2 C: 1.27 1.82 mm3 S: 1.17 1.53 mm3
4 Coronal (C) and sagittal (S) Coronal: skull vertex to feet. Sagittal: axial skeleton / 1000 17.50 — 90 1 C: 1.58 1.97 mm3 S: 1.38 1.75 mm3
4 Axial Skull vertex to feet 0, 500, 1000 s/mm2 1400 66 180 90 2 5 4.97 mm3
DWIBS, diffusion-weighted imaging with background body signal suppression; FA, flip angle; NSA, number of signals acquired; STIR, short tau inversion recovery; TE, echo time; TI, inversion time; TR, repetition time; TSE, turbo spine echo.
neoplastic lesions. According to previously reported qualitative analysis of DWI (20), the primary bone tumor (either as focal lesion or diffuse involvement) appeared on DWIBS sequences as an area with high signal intensity, as compared to normal bone marrow; the response to chemotherapy was established as a partial/slight or complete decrease of signal intensity after treatment. Myelomatous lesions were defined as hypointense on T1 images and demonstrated an increase of signal intensity on STIR sequences (19). Afterward, DWI images were qualitatively analyzed. ADC maps were automatically generated reckoning with all the three different b values of the DWIBS sequences. For proper quantification, only well-defined lesions with diameter >2 cm or diffusely involving a skeletal segment (eg, vertebral body) were evaluated and considered in our analysis as target lesions (up to two per patient). Hence, for lesions recorded as positive on DWI, the ones with the lowest signal intensity on the corresponding ADC maps were evaluated. The mean ADC values of each lesion were calculated both in responders and nonresponders, using a dedicated workstation (ViewForum; Philips Medical Systems). For this purpose, a region of interest (ROI) was manually drawn on the ADC maps along the contours of the solid focal lesion, covering its entire volume on a single slice. In patients with diffuse involvement, an ROI was drawn within the bone region with altered signal intensity (ie, vertebral body): the size was not
I
n the past decades, magnetic resonance imaging (MRI) has shown a high diagnostic and prognostic value in monoclonal plasma cell disorders (1–4). It has also been included as additional imaging tool for diagnosis and staging protocols of multiple myeloma (MM) (2,4), but, to date,
Acad Radiol 2015; 22:1163–1171 From the School of Medicine, University of Milano-Bicocca, Via Pergolesi 33, 20900 Monza, Italy (P.A.B., D.I., A.C., V.B., C.T.F., S.S.); Department of Diagnostic Radiology, H. San Gerardo, Monza, Italy (P.A.B., D.I., A.C., V.B., C.T.F., S.S.); and Department of Health Sciences, Tecnomed Foundation, University of Milano-Bicocca, Monza, Italy (P.A.B.). Received November 23, 2014; accepted May 27, 2015. Address correspondence to: P.A.B. e-mail: [email protected] ªAUR, 2015 http://dx.doi.org/10.1016/j.acra.2015.05.011
there is insufficient evidence to recommend routine MRI after treatment because most of the findings are not specific (1,2,5). In MM follow-up, the role of imaging is still limited, relying on skeletal survey although bone marrow aspirate and laboratory parameters (serum and urinary M-protein measurements) are the mainstay of the evaluation of response to chemotherapy. According to the most recent guidelines, plain films should be repeated after clinical or laboratory evidence of disease progression (2,5–7), but on conventional radiographs, lytic bone lesions rarely demonstrate signs of healing. Moreover, the evidence of new vertebral compressions may not indicate disease progression, occurring even after an effective therapy (1,5). Therefore, an accurate imaging technique able to detect treatment-related changes 1163
BONAFFINI ET AL
could increase the ability to monitor MM patients, even those with reduced or absent M-protein secretion. Diffusion-weighted imaging (DWI) is an additional and noninvasive tool that allows visualization and measurement of microscopic Brownian motion (diffusivity) of water molecules in the tissues. Thus, MRI coupled with DWI can provide within a single examination, morphological details along with functional information, related to tissue cellularity, to the cell membrane integrity and to intracellular/extracellular space–volume ratio (8–11). The introduction of DWI with background body signal suppression (DWIBS) technique (12) allows the acquisition, during free breathing, of volumetric and thin-sliced diffusion-weighted images of the whole body (8), making it applicable even in MM patients. Along with visual and qualitative analysis, diffusion can be also quantified in biological tissues by means of the so-called apparent diffusion coefficient (ADC) (8,9). Several studies have shown the ability of DWI in the detection and characterization of different pathologic processes (acute ischemic stroke, inflammatory lesions, malignant tumors) and in the differentiation of postoperative changes from residual active tumor or recurrence (11,13–16). DWI role in the evaluation of response to treatment has to be still clearly defined, and to knowledge, there are few studies that evaluated it in MM (7,17,18). Nevertheless, usefulness of DWI in this issue is expected because of high cellular load replacing hematopoietic bone marrow that characterized MM (7,19,20). The purpose of this study was to assess the value of quantitative analysis of ADC maps, obtained by DWIBS sequences, in the evaluation of response to chemotherapy in MM patients. MATERIALS AND METHODS Study Population
Institutional ethical committee approval was obtained for this retrospective analysis of a total of 28 MRI studies performed in 14 patients (seven women; average age, 60 years; age range, 40–76 years). The main inclusion criteria were patients referred to our department with histologic diagnosis of MM by means of bone marrow aspirate, undergoing whole-body MRI (WB-MRI) examination, and who were eligible to treatment. Patients with other malignant tumors (ie, lung cancer) and who were not able to properly complete the whole MRI study were excluded. Every patient underwent a WB-MRI study both for initial staging (MR1) and on the average 7 months (range, 2–17) after chemotherapy treatment (MR2). Informed consent for MRI was obtained from all the patients. All the included patients were treated with standard protocols, which included chemotherapy drugs (bortezomib, thalidomide, vincristine, doxorubicin, melphalan), alone or combined with the administration of dexamethasone. 1164
Academic Radiology, Vol 22, No 9, September 2015
MRI Technique
Every WB-MRI examination was performed on a 1.5T scanner (Achieva; Philips Medical Systems, Best, The Netherlands), using a built-in body coil combined with stepping table technique, as supplied by the manufacturer and during free but regular breathing. No patient preparation was required, and no contrast material was injected. The patients were positioned in the magnet supine and head first. After the acquisition of survey images of the whole body (range, 5 to 7, depending on patient’s height), the WB-MRI protocol consisted of the following sequences (Table 1): T2 short tau inversion recovery (STIR; turbo factor, 30; repetition time [TR], 1400; echo time [TE], 64; inversion time [TI], 165; flip angle [FA], 90 ; number of signals acquired [NSA], 2) and T1 turbo spine echo (TSE; TR, 1000; TE, 17.50; FA, 90 ; NSA, 1) sequences, acquired on the coronal plane from the skull vertex to the feet (T1: voxel size 1.27 1.82 mm3; STIR: voxel size 1.58 1.97 mm3). The field of view (FOV) was manually adapted to the major circumference of the patient. The same sequences were also acquired on the sagittal plane covering the whole axial skeleton (T1: voxel size 1.17 1.53 mm3; STIR: voxel size 1.38 1.75 mm3). On the same coverage area, DWIBS sequences were acquired on the axial plane to avoid distortion artifacts, using three different b values (b = 0, b = 500 e b = 1000 s/mm2) and with the following parameters: echo-planar imaging (EPI), TR, 1400; TE, 66; FA, 90 ; NSA, 2; voxel size 5 4.97 mm3. To obtain fat suppression, DWIBS was combined with an STIR prepulse. All the study sequences were acquired during free breathing, with slice thickness of 4and 1-mm gap. The total acquisition time was about 60–75 minutes, depending on patient’s height. At the end of the study, every imaged district was merged using software integrated in the scanner, generating coronal whole-body T1, T2 STIR, and DWIBS reconstructions (Fig. 1). Moreover, inverting the gray scale of DWIBS images on reporting workstation, ‘‘PET-like’’ images were available for radiologists at the time of evaluation of examination. Patients and Imaging Analysis
At the time of diagnosis and during follow-up, the percentage of plasma cells by the means of the bone marrow aspirate was evaluated, and the response to treatment was established as a reduction of bone marrow infiltration (6). When histologic analysis was not available at our Institution’s electronic database (2 of 14 patients, 1 responder), the response to treatment was established according to the clinical evaluation by medical records. At the first examination, axial DWIBS sequences were compared with standard T1 and T2 STIR images in five different anatomic districts (skull, spine, sternum and ribs, pelvis, upper and lower extremities) to detect the primary
Academic Radiology, Vol 22, No 9, September 2015
ADC MAPS FOR RESPONSE TO CHEMOTHERAPY IN MM
TABLE 1. Whole-Body Magnetic Resonance Imaging Protocol (1.5T), Acquired with Patient Positioned Supine and Using Built-In Body Coil Combined with Stepping Table Technique Sequence Slice thickness, mm Plane of acquisition Range of scan b values TR TE TI FA NSA Voxel size
T2 STIR
T1 TSE
DWIBS
4 Coronal (C) and sagittal (S) Coronal: skull vertex to feet. Sagittal: axial skeleton / 1400 64 165 90 2 C: 1.27 1.82 mm3 S: 1.17 1.53 mm3
4 Coronal (C) and sagittal (S) Coronal: skull vertex to feet. Sagittal: axial skeleton / 1000 17.50 — 90 1 C: 1.58 1.97 mm3 S: 1.38 1.75 mm3
4 Axial Skull vertex to feet 0, 500, 1000 s/mm2 1400 66 180 90 2 5 4.97 mm3
DWIBS, diffusion-weighted imaging with background body signal suppression; FA, flip angle; NSA, number of signals acquired; STIR, short tau inversion recovery; TE, echo time; TI, inversion time; TR, repetition time; TSE, turbo spine echo.
neoplastic lesions. According to previously reported qualitative analysis of DWI (20), the primary bone tumor (either as focal lesion or diffuse involvement) appeared on DWIBS sequences as an area with high signal intensity, as compared to normal bone marrow; the response to chemotherapy was established as a partial/slight or complete decrease of signal intensity after treatment. Myelomatous lesions were defined as hypointense on T1 images and demonstrated an increase of signal intensity on STIR sequences (19). Afterward, DWI images were qualitatively analyzed. ADC maps were automatically generated reckoning with all the three different b values of the DWIBS sequences. For proper quantification, only well-defined lesions with diameter >2 cm or diffusely involving a skeletal segment (eg, vertebral body) were evaluated and considered in our analysis as target lesions (up to two per patient). Hence, for lesions recorded as positive on DWI, the ones with the lowest signal intensity on the corresponding ADC maps were evaluated. The mean ADC values of each lesion were calculated both in responders and nonresponders, using a dedicated workstation (ViewForum; Philips Medical Systems). For this purpose, a region of interest (ROI) was manually drawn on the ADC maps along the contours of the solid focal lesion, covering its entire volume on a single slice. In patients with diffuse involvement, an ROI was drawn within the bone region with altered signal intensity (ie, vertebral body): the size was not
