Eur J Nucl Med Mol Imaging DOI 10.1007/s00259-014-2756-0
ORIGINAL ARTICLE
The value of 18F-FDG PET/CT in the management of malignant peripheral nerve sheath tumors Benjapa Khiewvan & Homer A. Macapinlac & Dina Lev & Ian E. McCutcheon & John M. Slopis & Ghadah Al Sannaa & Wei Wei & Hubert H. Chuang
Received: 4 November 2013 / Accepted: 4 March 2014 # Springer-Verlag Berlin Heidelberg 2014
Abstract Purpose Our objective was to determine how positron emission tomography (PET)/CT had been used in the clinical treatment of malignant peripheral nerve sheath tumor (MPNST) patients at The University of Texas MD Anderson Cancer Center. Methods We reviewed a database of MPNST patients referred to MD Anderson Cancer Center during 1995–2011. We enrolled 47 patients who underwent PET/CT imaging. Disease stage was based on conventional imaging and PET/CT B. Khiewvan : H. A. Macapinlac : H. H. Chuang (*) Department of Nuclear Medicine, University of Texas MD Anderson Cancer Center, P.O. Box 301402, Unit 1483,1515 Holcombe Blvd., Houston, TX 77230-1402, USA e-mail: [email protected] B. Khiewvan Division of Nuclear Medicine, Department of Radiology, Faculty of Medicine Siriraj Hospital, Mahidol University, 2 Prannok Road Siriraj Bangkoknoi, Bangkok 10700, Thailand D. Lev : G. Al Sannaa Department of Cancer Biology, The University of Texas MD Anderson Cancer Center, Unit 0085, 1515 Holcombe Blvd., Houston, TX 77030, USA I. E. McCutcheon Department of Neurosurgery, The University of Texas MD Anderson Cancer Center, Unit 0442, 1400 Holcombe Blvd., Houston, TX 77030, USA J. M. Slopis Department of Neuro-Oncology, The University of Texas MD Anderson Cancer Center, Unit 0431, 1400 Holcombe Blvd., Houston, TX 77030, USA W. Wei Department of Biostatistics, The University of Texas MD Anderson Cancer Center, Unit 1411, 1515 Holcombe Blvd., Houston, TX 77030, USA
findings using National Comprehensive Cancer Network (NCCN) guidelines. Treatment strategies based on PET/CT and conventional imaging were determined by chart review. The maximum and mean standardized uptake values (SUVmax, SUVmean), metabolic tumor volume (MTV), total lesion glycolysis (TLG), change in SUVmax, change in MTV, and change in TLG were calculated from the PET/CT studies before and after treatment. Response prediction was based on imaging studies performed before and after therapy and categorized as positive or negative for residual tumor. Clinical outcome was determined from chart review. Results PET/CT was performed for staging in 16 patients, for restaging in 29 patients, and for surveillance in 2 patients. Of the patients, 88 % were correctly staged with PET/CT, whereas 75 % were correctly staged with conventional imaging. The sensitivity to detect local recurrence and distant metastasis at restaging was 100 and 100 % for PET/CT compared to 86 and 83 % for conventional imaging, respectively. PET/CT findings resulted in treatment changes in 31 % (5/16) and 14 % (4/29) of patients at staging and restaging, respectively. Recurrence, MTV, and TLG were prognostic factors for survival, whereas SUVmax and SUVmean were not predictive. For 21 patients who had imaging studies performed both before and after treatment, PET/CT was better at predicting outcome (overall survival, progression-free survival) than conventional imaging. A decreasing SUVmax ≥ 30 % and decrease in TLG and MTV were significant predictors for overall and progression-free survival. Conclusion PET/CT is valuable in MPNST management because of its high accuracy in staging and high sensitivity and accuracy in restaging as well as improvements in treatment planning. MTV from baseline staging studies is predictive of survival. Additionally, change in SUVmax, TLG, and MTV accurately predicted outcomes after treatment.
Eur J Nucl Med Mol Imaging
Keywords Malignant peripheral nerve sheath tumor . 18 F-FDG PET/CT . Treatment impact . Prognostic value
Introduction Malignant peripheral nerve sheath tumors (MPNSTs) are rare and aggressive tumors that account for 5–10 % of soft tissue sarcomas [1]. MPNSTs arise from peripheral nerves or cells associated with the nerve sheath, such as Schwann cells, fibroblasts, or perineural cells. Approximately 50 % of these tumors are the result of malignant transformation of plexiform neurofibromas in patients with neurofibromatosis type 1 (NF1) [2], but they may also occur sporadically. Presenting symptoms are a rapidly growing mass, pain, and/or nerve loss. The first line of treatment is surgery. Chemotherapy or radiation therapy can be used for adjuvant treatment. The prognosis for MPNST patients is generally poor: the risk of local recurrence is 38–45 %, and the risk of distant metastases is 40–82 % [3]. Many methods exist for the evaluation of MPNSTs. The main diagnostic imaging modalities are magnetic resonance imaging (MRI) and computed tomography (CT). These modalities provide information about tumor margins, heterogeneity, and invasion. Chest CT and bone scan are used to evaluate distant metastasis. Positron emission tomography (PET) is a noninvasive imaging modality that provides molecular and physiological information. Moreover, CT combined with PET allows better localization and improvement of sensitivity and specificity. 18 F-Labeled fluorodeoxyglucose (FDG) is mainly used in clinical oncological imaging and reflects glucose metabolism in the human body. 18F-FDG PET/CT is used in sarcoma treatment for tumor grading assessment, tumor biopsy guidance, staging, restaging, assessment of treatment response, and determining prognosis. 18F-FDG PET/CT has been shown to differentiate malignant from benign MPNSTs and identify malignant change in NF1 [4, 5]. However, the role of PET/CT in restaging, management change, and determining prognosis in MPNSTs is not well established. In this study, we sought to determine how effective PET/CT has been in the clinical treatment of MPNST patients at The University of Texas MD Anderson Cancer Center.
Materials and methods Patients We retrospectively reviewed a database of 267 MPNST patients referred to MD Anderson Cancer Center between 1995
and 2011 to identify patients who had undergone PET/CT imaging. Forty-seven patients included in the study had undergone conventional imaging with CT, MRI, and/or bone scan within 30 days of the PET/CT studies. Following National Comprehensive Cancer Network (NCCN) guidelines, staging was determined by the American Joint Committee on Cancer staging system (7th edition) for soft tissue sarcoma [6]. Clinical details, treatment strategies based on PET/CT and conventional imaging, and follow-up information were obtained by chart review. This study was approved by the Institutional Review Board of MD Anderson and was compliant with the Health Insurance Portability and Accountability Act. Informed consent was waived by the Institutional Review Board.
18
F-FDG PET/CT imaging
Patients received an intravenous injection of 185–370 MBq of 18 F-FDG after 6 h of fasting, provided their blood glucose levels were less than 150 mg/dl. Whole-body PET/CT image acquisition (Discovery ST, STe, or RX, GE Healthcare) commenced approximately 60–90 min after FDG injection in 2D or 3D modes. The acquisition time was 3–5 min per bed position, according to the patient’s weight. A noncontrastenhanced CT scan (140 kV, 120 mA, 3.7-mm axial slice placement, and 13.5-mm table speed) was used for attenuation correction and diagnostic purposes. Images were reconstructed using an ordered subset expectation maximization algorithm. PET/CT images were visually and semiquantitatively interpreted for abnormal radiotracer uptake above the background corresponding with abnormalities seen on the CT scan. Standardized uptake value (SUV) was calculated as follows:
SUV ¼
tissue radioactivity concentration ðBq=mLÞ injected activity ðBqÞ=body weight ðgÞ
The maximum standardized uptake value (SUVmax) was the highest metabolic activity and was used for semiquantitative analysis. However, SUVmax may not represent the status of the whole tumor. Metabolic tumor volume (MTV) was used to measure the total tumor volume segmented via threshold SUV [7]. MTV was calculated using a 45 % threshold from SUVmax in this study to provide a high tumor to background ratio and reduce partial volume effects [8, 9]. Total lesion glycolysis (TLG) was calculated as the product of MTV multiplied by SUVmean. In this study, we determined whether the cutoff values could be used as predictors of survival, using median values of SUVmax, SUVmean, MTV, and TLG as well as a percentage change in SUV max (ΔSUV max), MTV (ΔMTV), and TLG (ΔTLG).
Eur J Nucl Med Mol Imaging
We calculated percentage change in SUVmax, MTV, and TLG as follows:
SUVmax2 −SUVmax1 100 SUVmax1 MTV2 −MTV1 100 ΔMTV ¼ MTV1 f½ðSUVmean2 Þ ðtumor volume2 Þ−½ðSUVmean1 Þ ðtumor volume1 Þg 100 ΔTLG ¼ ðSUVmeanÞ1 ðtumor volumeÞ1
ΔSUVmax ¼
PET/CT scan and conventional imaging were performed after initial treatment in 21 MPNST patients to evaluate treatment response. The mean duration time between treatment and follow-up imaging studies was 5.5 months (range 1– 25 months). Patients were treated as follows [6]: Patients with stage II disease were treated with surgery, chemoradiation, or surgery plus preoperative chemoradiation or chemotherapy. Patients with stage III disease were treated with surgery plus postoperative adjuvant chemoradiation or surgery plus preoperative chemotherapy and postoperative adjuvant chemotherapy or chemoradiation. Patients with stage IV disease were treated with chemotherapy or chemotherapy plus palliative radiation therapy. Patients with tumor recurrence were treated with radiotherapy, surgery, or surgery plus adjuvant chemotherapy or radiation therapy. Patients with distant recurrence were treated with chemotherapy. Follow-up PET/CT scans and conventional imaging after treatment were categorized as either positive or negative for active disease based upon a review of the reports. The PET/CT and conventional imaging response after treatment was analyzed for survival prediction. Imaging classification We analyzed the findings from conventional imaging and PET/CT images for the accuracy for detecting local recurrence and distant metastasis based on patient and anatomical region. Results were defined as true-positive (TP), true-negative (TN), false-positive (FP), and false-negative (FN) according to pathological results or subsequent imaging within 6 months. Twenty-two patients were analyzed because the patients’ follow-up time after imaging studies was more than 6 months. For detecting local recurrence, 14 of the 22 patients (63.6 %) were classified based on pathological results, while 8 of the 22 patients (36.4 %) were classified based on clinical follow-up data and follow-up imaging. For detecting distant recurrence, the pathological results was available in three of six distant recurrent cases (50 %), while the other three patients (50 %) were determined to have distant recurrence based on clinical
follow-up data and follow-up imaging. The remaining 16 patients did not show evidence of distant recurrence during the 6 months of follow-up after the restaging PET/CT scan. Impact of PET/CT on treatment Treatment planning was determined by conventional imaging and PET/CT findings. The impact of PET/CT on patient treatment was calculated as the percentage change of treatment determined by conventional imaging vs PET/CT study. Clinical outcomes Overall survival (OS) and progression-free survival (PFS) durations were recorded as clinical outcome measures. OS was defined as the number of months from the initial treatment after the PET/CT date until death or the last documented follow-up date. PFS was defined as the number of months after the initiation of treatment after the PET/CT date until disease progression, recurrence, or death. These data were obtained through chart review. OS and PFS were used to analyze patient prognosis. Statistical analysis All categorical data were calculated by using counts and percentages. Continuous variables are expressed as mean± standard deviation, median, and range. The accuracy of PET/ CT vs conventional imaging at staging was obtained. Patientbased and region-based analyses of PET/CT vs conventional imaging to determine local recurrence and metastasis were compared. Sensitivity, specificity, and accuracy were calculated. Comparison between PET/CT and conventional imaging modalities with respect to accuracy was performed using the McNemar test. The Kaplan-Meier method was used to estimate patient survival endpoints (OS and PFS). For survival endpoints, medians for imaging parameters (SUVmax, SUVmean, MTV,
Eur J Nucl Med Mol Imaging Table 1 Comparison between conventional and PET/CT staging in MPNST patients Patient no.
Final staging
Conventional staging
PET staging
Stage migration
1 2 3 4 5 6 7 8 9
I I II II II II II II II
0a IV 0a II II II II II II
I I II II II II II II III
Upstage Downstage Upstage No No No No No Upstage
10 11 12 13 14 15 16
II III III III IV IV IV
II III III III III IV IV
III III III III IV IV IV
Upstage No No No Upstage No No Upstage=5/16 (31.3 %) Downstage=1/16 (6.3 %)
especially highly skewed data. A data-driven cutoff point is not appropriate for small sample sizes which cannot be validated using cross-validation or an independent test set. Comparisons of survival endpoints between patient groups and the associations between OS and covariates of interest (age, sex, NF1 syndrome, stage, SUVmax, SUVmean, MTV, TLG, and recurrence) were carried out using the log-rank test. The response predictions of PET/CT and conventional imaging were analyzed using the Kaplan-Meier method. All tests were two-sided and p values of 0.05 or less were considered statistically significant. Statistical analysis was carried out using SAS version 9 (SAS Institute, Cary, NC, USA).
Results Patient characteristics
and TLG), ΔSUVmax, ΔMTV, and ΔTLG were used to dichotomize patients into low and high groups. The median was used as a cutoff point because it is robust to data distribution,
There were 47 patients (25 men and 22 women, age range 12– 76 years, median age 38 years) with confirmed pathological diagnosis of MPNST, referred to MD Anderson, who had PET/ CT as part of their management. The patients were white (n=32), African-American (n=6), Hispanic (n=7), and Asian (n=2). The patients had NF1 (n=25), schwannomatosis (n=3), or neither (i.e., they were sporadic, n=19). They presented with mass (n=25), pain (n=8), neurological disorders (n=2), mass with pain (n=4), mass with neurological disorders (n=3), and pain with neurological disorders (n=5). PET/CT scans were performed for staging (n=16), restaging (n=29), and surveillance (n=2). The
Fig. 1 a, b CT images showed MPNST at paramidline region with lung metastasis (white arrow). c Bone scan showed bony metastasis (stage IV). d–g PET/CT images confirmed primary tumor but neither evidence of
lung (yellow arrow) nor bony metastases (stage I). The pathology and follow-up CT scans showed stage I cancer. The treatment changed from palliative to curative surgery
a
No evidence of MPNST
Eur J Nucl Med Mol Imaging Table 2 Patient-based results from PET/CT vs conventional imaging in restaging
Diagnostic modality
No. of patients
Sensitivity
Specificity
Accuracy
FN
FP
TN
TP
No./total (%)
No./total (%)
No./total (%)
0 2
2 1
6 7
14 12
14/14 (100) 12/14 (86)
6/8 (75) 7/8 (88)
20/22 (91) 19/22 (86)
0 1
0 2
16 14
6 5
6/6 (100) 5/6 (83)
16/16 (100) 14/16 (88)
22/22 (100) 19/22 (86)
Local recurrence
FN false-negative, FP false-positive, TN true-negative, TP truepositive
PET/CT Conventional imaging Distant metastasis PET/CT Conventional imaging
follow-up time of the 47 patients after the first visit ranged from 7 to 216 months (median 33 months). Accuracy of PET/CT vs conventional imaging Stage assignments, based on conventional imaging findings, PET/CT findings, and final clinical stage, are shown for 16 patients (Table 1). PET/CT and conventional imaging agreed on staging in 62.5 %, but differed in 37.5 % of cases. PET/CT correctly identified the stage in 87.5 % (14/ 16), whereas conventional imaging correctly identified the stage in 75.0 % (12/16), but not with a statistically significant difference (p=0.41). PET/CT incorrectly identified tumor in the lymph nodes of two patients; malignancy was not proven on pathology analysis. Conventional imaging missed lung metastasis in one patient and a primary tumor in two patients and detected FP lung and bone lesions in one patient (Fig. 1). Among these 16 patients, PET/CT upstaged the diagnosis in 5 cases (31.3 %) and downstaged it in 1 case (6.3 %). Of the 29 restaging cases, 7 cases were excluded owing to a follow-up time after imaging studies of less than 6 months. There was recurrent disease in 14 patients. For local
recurrence, PET/CT results were described as TP in 14 patients (sensitivity 100 %), TN in 6 patients (specificity 75 %), and FP in 2 patients; there was no FN case (Table 2). Conventional imaging results showed TP in 12 patients (sensitivity 86 %), FN in 2 patients, TN in 7 patients (specificity 88 %), and FP in 1 patient. For distant metastasis, PET/CT results were TP in 6 patients (sensitivity 100 %) and TN in 16 patients (specificity 100 %); there were no FN and FP cases. Conventional imaging was TP in 5 patients (sensitivity 83 %), FN in 1 patient, TN in 14 patients (specificity 88 %), and FP in 2 patients. The accuracy of PET/CT was higher than that of conventional imaging (91 vs 86 % for local recurrence and 100 vs 86 % for distant metastasis), but the differences were not statistically significant (p=0.1). Impact on treatment planning PET/CT findings resulted in a change in the treatment plan in 31.3 % (5/16) of staging cases and 13.8 % (4/29) of restaging cases (Table 3). There was no treatment change in surveillance cases. PET/CT had a high impact owing to a change from palliative to curative treatment in four patients (8.5 %). Figure 2 displays a sample case with a change in treatment
Table 3 Change in patient management by staging and restaging PET/CT Treatment planning by conventional imaging Staging Surgery Preoperative RT, surgery Chemotherapy Palliative chemoradiation Observation Restaging Surgery
Treatment planning by PET/CT
n
Reason for treatment change
Neoadjuvant therapy, surgery Chemoradiation, surgery Chemoradiation Surgery
1 1 1 1
Surgery
1
PET/CT upstaged from stage II to III PET/CT upstaged from stage II to III Additional bony metastatic lesion at left knee from PET/CT Lung and bone metastases on conventional imaging were shown to be benign by PET/CT Suspicion of malignant lesion from PET/CT
Surgery at different location
1
Chemotherapy Neoadjuvant therapy, surgery 3 Impact on treatment planning: staging 5/16 (31.3 %), restaging 4/29 (13.8 %) RT radiation therapy
False-positive tumor recurrence not shown on PET/CT. Additional avid lesion in abdomen seen on PET/CT PET/CT showed single lesion of disease
Eur J Nucl Med Mol Imaging Fig. 2 a, b CT showed primary tumor and lung metastasis (white arrow). c, d PET/CT showed primary tumor with benign lung lesion (yellow arrow). The treatment plan changed from systemic to local therapy. Pathology of the lung lesion was pulmonary hemangioma
from palliative to curative intent. Due to upstaging, PET/CT caused clinicians to give more adjuvant treatment in two patients and it detected additional metastatic disease in one
patient (6.4 %). PET/CT changed treatment from observation to surgery in one patient (2.1 %). Moreover, PET/CT changed the location of surgery in one patient (2.1 %). In this patient,
Fig. 3 Staging SUVmax (a) and restaging SUVmax (d) did not predict OS. Staging MTVand TLG (b, c) and restaging MTVand TLG (e, f) predicted OS
Eur J Nucl Med Mol Imaging Table 4 Prognostic factors for survival by univariate analysis (staging) Factor Age ≥40 years
ORIGINAL ARTICLE
The value of 18F-FDG PET/CT in the management of malignant peripheral nerve sheath tumors Benjapa Khiewvan & Homer A. Macapinlac & Dina Lev & Ian E. McCutcheon & John M. Slopis & Ghadah Al Sannaa & Wei Wei & Hubert H. Chuang
Received: 4 November 2013 / Accepted: 4 March 2014 # Springer-Verlag Berlin Heidelberg 2014
Abstract Purpose Our objective was to determine how positron emission tomography (PET)/CT had been used in the clinical treatment of malignant peripheral nerve sheath tumor (MPNST) patients at The University of Texas MD Anderson Cancer Center. Methods We reviewed a database of MPNST patients referred to MD Anderson Cancer Center during 1995–2011. We enrolled 47 patients who underwent PET/CT imaging. Disease stage was based on conventional imaging and PET/CT B. Khiewvan : H. A. Macapinlac : H. H. Chuang (*) Department of Nuclear Medicine, University of Texas MD Anderson Cancer Center, P.O. Box 301402, Unit 1483,1515 Holcombe Blvd., Houston, TX 77230-1402, USA e-mail: [email protected] B. Khiewvan Division of Nuclear Medicine, Department of Radiology, Faculty of Medicine Siriraj Hospital, Mahidol University, 2 Prannok Road Siriraj Bangkoknoi, Bangkok 10700, Thailand D. Lev : G. Al Sannaa Department of Cancer Biology, The University of Texas MD Anderson Cancer Center, Unit 0085, 1515 Holcombe Blvd., Houston, TX 77030, USA I. E. McCutcheon Department of Neurosurgery, The University of Texas MD Anderson Cancer Center, Unit 0442, 1400 Holcombe Blvd., Houston, TX 77030, USA J. M. Slopis Department of Neuro-Oncology, The University of Texas MD Anderson Cancer Center, Unit 0431, 1400 Holcombe Blvd., Houston, TX 77030, USA W. Wei Department of Biostatistics, The University of Texas MD Anderson Cancer Center, Unit 1411, 1515 Holcombe Blvd., Houston, TX 77030, USA
findings using National Comprehensive Cancer Network (NCCN) guidelines. Treatment strategies based on PET/CT and conventional imaging were determined by chart review. The maximum and mean standardized uptake values (SUVmax, SUVmean), metabolic tumor volume (MTV), total lesion glycolysis (TLG), change in SUVmax, change in MTV, and change in TLG were calculated from the PET/CT studies before and after treatment. Response prediction was based on imaging studies performed before and after therapy and categorized as positive or negative for residual tumor. Clinical outcome was determined from chart review. Results PET/CT was performed for staging in 16 patients, for restaging in 29 patients, and for surveillance in 2 patients. Of the patients, 88 % were correctly staged with PET/CT, whereas 75 % were correctly staged with conventional imaging. The sensitivity to detect local recurrence and distant metastasis at restaging was 100 and 100 % for PET/CT compared to 86 and 83 % for conventional imaging, respectively. PET/CT findings resulted in treatment changes in 31 % (5/16) and 14 % (4/29) of patients at staging and restaging, respectively. Recurrence, MTV, and TLG were prognostic factors for survival, whereas SUVmax and SUVmean were not predictive. For 21 patients who had imaging studies performed both before and after treatment, PET/CT was better at predicting outcome (overall survival, progression-free survival) than conventional imaging. A decreasing SUVmax ≥ 30 % and decrease in TLG and MTV were significant predictors for overall and progression-free survival. Conclusion PET/CT is valuable in MPNST management because of its high accuracy in staging and high sensitivity and accuracy in restaging as well as improvements in treatment planning. MTV from baseline staging studies is predictive of survival. Additionally, change in SUVmax, TLG, and MTV accurately predicted outcomes after treatment.
Eur J Nucl Med Mol Imaging
Keywords Malignant peripheral nerve sheath tumor . 18 F-FDG PET/CT . Treatment impact . Prognostic value
Introduction Malignant peripheral nerve sheath tumors (MPNSTs) are rare and aggressive tumors that account for 5–10 % of soft tissue sarcomas [1]. MPNSTs arise from peripheral nerves or cells associated with the nerve sheath, such as Schwann cells, fibroblasts, or perineural cells. Approximately 50 % of these tumors are the result of malignant transformation of plexiform neurofibromas in patients with neurofibromatosis type 1 (NF1) [2], but they may also occur sporadically. Presenting symptoms are a rapidly growing mass, pain, and/or nerve loss. The first line of treatment is surgery. Chemotherapy or radiation therapy can be used for adjuvant treatment. The prognosis for MPNST patients is generally poor: the risk of local recurrence is 38–45 %, and the risk of distant metastases is 40–82 % [3]. Many methods exist for the evaluation of MPNSTs. The main diagnostic imaging modalities are magnetic resonance imaging (MRI) and computed tomography (CT). These modalities provide information about tumor margins, heterogeneity, and invasion. Chest CT and bone scan are used to evaluate distant metastasis. Positron emission tomography (PET) is a noninvasive imaging modality that provides molecular and physiological information. Moreover, CT combined with PET allows better localization and improvement of sensitivity and specificity. 18 F-Labeled fluorodeoxyglucose (FDG) is mainly used in clinical oncological imaging and reflects glucose metabolism in the human body. 18F-FDG PET/CT is used in sarcoma treatment for tumor grading assessment, tumor biopsy guidance, staging, restaging, assessment of treatment response, and determining prognosis. 18F-FDG PET/CT has been shown to differentiate malignant from benign MPNSTs and identify malignant change in NF1 [4, 5]. However, the role of PET/CT in restaging, management change, and determining prognosis in MPNSTs is not well established. In this study, we sought to determine how effective PET/CT has been in the clinical treatment of MPNST patients at The University of Texas MD Anderson Cancer Center.
Materials and methods Patients We retrospectively reviewed a database of 267 MPNST patients referred to MD Anderson Cancer Center between 1995
and 2011 to identify patients who had undergone PET/CT imaging. Forty-seven patients included in the study had undergone conventional imaging with CT, MRI, and/or bone scan within 30 days of the PET/CT studies. Following National Comprehensive Cancer Network (NCCN) guidelines, staging was determined by the American Joint Committee on Cancer staging system (7th edition) for soft tissue sarcoma [6]. Clinical details, treatment strategies based on PET/CT and conventional imaging, and follow-up information were obtained by chart review. This study was approved by the Institutional Review Board of MD Anderson and was compliant with the Health Insurance Portability and Accountability Act. Informed consent was waived by the Institutional Review Board.
18
F-FDG PET/CT imaging
Patients received an intravenous injection of 185–370 MBq of 18 F-FDG after 6 h of fasting, provided their blood glucose levels were less than 150 mg/dl. Whole-body PET/CT image acquisition (Discovery ST, STe, or RX, GE Healthcare) commenced approximately 60–90 min after FDG injection in 2D or 3D modes. The acquisition time was 3–5 min per bed position, according to the patient’s weight. A noncontrastenhanced CT scan (140 kV, 120 mA, 3.7-mm axial slice placement, and 13.5-mm table speed) was used for attenuation correction and diagnostic purposes. Images were reconstructed using an ordered subset expectation maximization algorithm. PET/CT images were visually and semiquantitatively interpreted for abnormal radiotracer uptake above the background corresponding with abnormalities seen on the CT scan. Standardized uptake value (SUV) was calculated as follows:
SUV ¼
tissue radioactivity concentration ðBq=mLÞ injected activity ðBqÞ=body weight ðgÞ
The maximum standardized uptake value (SUVmax) was the highest metabolic activity and was used for semiquantitative analysis. However, SUVmax may not represent the status of the whole tumor. Metabolic tumor volume (MTV) was used to measure the total tumor volume segmented via threshold SUV [7]. MTV was calculated using a 45 % threshold from SUVmax in this study to provide a high tumor to background ratio and reduce partial volume effects [8, 9]. Total lesion glycolysis (TLG) was calculated as the product of MTV multiplied by SUVmean. In this study, we determined whether the cutoff values could be used as predictors of survival, using median values of SUVmax, SUVmean, MTV, and TLG as well as a percentage change in SUV max (ΔSUV max), MTV (ΔMTV), and TLG (ΔTLG).
Eur J Nucl Med Mol Imaging
We calculated percentage change in SUVmax, MTV, and TLG as follows:
SUVmax2 −SUVmax1 100 SUVmax1 MTV2 −MTV1 100 ΔMTV ¼ MTV1 f½ðSUVmean2 Þ ðtumor volume2 Þ−½ðSUVmean1 Þ ðtumor volume1 Þg 100 ΔTLG ¼ ðSUVmeanÞ1 ðtumor volumeÞ1
ΔSUVmax ¼
PET/CT scan and conventional imaging were performed after initial treatment in 21 MPNST patients to evaluate treatment response. The mean duration time between treatment and follow-up imaging studies was 5.5 months (range 1– 25 months). Patients were treated as follows [6]: Patients with stage II disease were treated with surgery, chemoradiation, or surgery plus preoperative chemoradiation or chemotherapy. Patients with stage III disease were treated with surgery plus postoperative adjuvant chemoradiation or surgery plus preoperative chemotherapy and postoperative adjuvant chemotherapy or chemoradiation. Patients with stage IV disease were treated with chemotherapy or chemotherapy plus palliative radiation therapy. Patients with tumor recurrence were treated with radiotherapy, surgery, or surgery plus adjuvant chemotherapy or radiation therapy. Patients with distant recurrence were treated with chemotherapy. Follow-up PET/CT scans and conventional imaging after treatment were categorized as either positive or negative for active disease based upon a review of the reports. The PET/CT and conventional imaging response after treatment was analyzed for survival prediction. Imaging classification We analyzed the findings from conventional imaging and PET/CT images for the accuracy for detecting local recurrence and distant metastasis based on patient and anatomical region. Results were defined as true-positive (TP), true-negative (TN), false-positive (FP), and false-negative (FN) according to pathological results or subsequent imaging within 6 months. Twenty-two patients were analyzed because the patients’ follow-up time after imaging studies was more than 6 months. For detecting local recurrence, 14 of the 22 patients (63.6 %) were classified based on pathological results, while 8 of the 22 patients (36.4 %) were classified based on clinical follow-up data and follow-up imaging. For detecting distant recurrence, the pathological results was available in three of six distant recurrent cases (50 %), while the other three patients (50 %) were determined to have distant recurrence based on clinical
follow-up data and follow-up imaging. The remaining 16 patients did not show evidence of distant recurrence during the 6 months of follow-up after the restaging PET/CT scan. Impact of PET/CT on treatment Treatment planning was determined by conventional imaging and PET/CT findings. The impact of PET/CT on patient treatment was calculated as the percentage change of treatment determined by conventional imaging vs PET/CT study. Clinical outcomes Overall survival (OS) and progression-free survival (PFS) durations were recorded as clinical outcome measures. OS was defined as the number of months from the initial treatment after the PET/CT date until death or the last documented follow-up date. PFS was defined as the number of months after the initiation of treatment after the PET/CT date until disease progression, recurrence, or death. These data were obtained through chart review. OS and PFS were used to analyze patient prognosis. Statistical analysis All categorical data were calculated by using counts and percentages. Continuous variables are expressed as mean± standard deviation, median, and range. The accuracy of PET/ CT vs conventional imaging at staging was obtained. Patientbased and region-based analyses of PET/CT vs conventional imaging to determine local recurrence and metastasis were compared. Sensitivity, specificity, and accuracy were calculated. Comparison between PET/CT and conventional imaging modalities with respect to accuracy was performed using the McNemar test. The Kaplan-Meier method was used to estimate patient survival endpoints (OS and PFS). For survival endpoints, medians for imaging parameters (SUVmax, SUVmean, MTV,
Eur J Nucl Med Mol Imaging Table 1 Comparison between conventional and PET/CT staging in MPNST patients Patient no.
Final staging
Conventional staging
PET staging
Stage migration
1 2 3 4 5 6 7 8 9
I I II II II II II II II
0a IV 0a II II II II II II
I I II II II II II II III
Upstage Downstage Upstage No No No No No Upstage
10 11 12 13 14 15 16
II III III III IV IV IV
II III III III III IV IV
III III III III IV IV IV
Upstage No No No Upstage No No Upstage=5/16 (31.3 %) Downstage=1/16 (6.3 %)
especially highly skewed data. A data-driven cutoff point is not appropriate for small sample sizes which cannot be validated using cross-validation or an independent test set. Comparisons of survival endpoints between patient groups and the associations between OS and covariates of interest (age, sex, NF1 syndrome, stage, SUVmax, SUVmean, MTV, TLG, and recurrence) were carried out using the log-rank test. The response predictions of PET/CT and conventional imaging were analyzed using the Kaplan-Meier method. All tests were two-sided and p values of 0.05 or less were considered statistically significant. Statistical analysis was carried out using SAS version 9 (SAS Institute, Cary, NC, USA).
Results Patient characteristics
and TLG), ΔSUVmax, ΔMTV, and ΔTLG were used to dichotomize patients into low and high groups. The median was used as a cutoff point because it is robust to data distribution,
There were 47 patients (25 men and 22 women, age range 12– 76 years, median age 38 years) with confirmed pathological diagnosis of MPNST, referred to MD Anderson, who had PET/ CT as part of their management. The patients were white (n=32), African-American (n=6), Hispanic (n=7), and Asian (n=2). The patients had NF1 (n=25), schwannomatosis (n=3), or neither (i.e., they were sporadic, n=19). They presented with mass (n=25), pain (n=8), neurological disorders (n=2), mass with pain (n=4), mass with neurological disorders (n=3), and pain with neurological disorders (n=5). PET/CT scans were performed for staging (n=16), restaging (n=29), and surveillance (n=2). The
Fig. 1 a, b CT images showed MPNST at paramidline region with lung metastasis (white arrow). c Bone scan showed bony metastasis (stage IV). d–g PET/CT images confirmed primary tumor but neither evidence of
lung (yellow arrow) nor bony metastases (stage I). The pathology and follow-up CT scans showed stage I cancer. The treatment changed from palliative to curative surgery
a
No evidence of MPNST
Eur J Nucl Med Mol Imaging Table 2 Patient-based results from PET/CT vs conventional imaging in restaging
Diagnostic modality
No. of patients
Sensitivity
Specificity
Accuracy
FN
FP
TN
TP
No./total (%)
No./total (%)
No./total (%)
0 2
2 1
6 7
14 12
14/14 (100) 12/14 (86)
6/8 (75) 7/8 (88)
20/22 (91) 19/22 (86)
0 1
0 2
16 14
6 5
6/6 (100) 5/6 (83)
16/16 (100) 14/16 (88)
22/22 (100) 19/22 (86)
Local recurrence
FN false-negative, FP false-positive, TN true-negative, TP truepositive
PET/CT Conventional imaging Distant metastasis PET/CT Conventional imaging
follow-up time of the 47 patients after the first visit ranged from 7 to 216 months (median 33 months). Accuracy of PET/CT vs conventional imaging Stage assignments, based on conventional imaging findings, PET/CT findings, and final clinical stage, are shown for 16 patients (Table 1). PET/CT and conventional imaging agreed on staging in 62.5 %, but differed in 37.5 % of cases. PET/CT correctly identified the stage in 87.5 % (14/ 16), whereas conventional imaging correctly identified the stage in 75.0 % (12/16), but not with a statistically significant difference (p=0.41). PET/CT incorrectly identified tumor in the lymph nodes of two patients; malignancy was not proven on pathology analysis. Conventional imaging missed lung metastasis in one patient and a primary tumor in two patients and detected FP lung and bone lesions in one patient (Fig. 1). Among these 16 patients, PET/CT upstaged the diagnosis in 5 cases (31.3 %) and downstaged it in 1 case (6.3 %). Of the 29 restaging cases, 7 cases were excluded owing to a follow-up time after imaging studies of less than 6 months. There was recurrent disease in 14 patients. For local
recurrence, PET/CT results were described as TP in 14 patients (sensitivity 100 %), TN in 6 patients (specificity 75 %), and FP in 2 patients; there was no FN case (Table 2). Conventional imaging results showed TP in 12 patients (sensitivity 86 %), FN in 2 patients, TN in 7 patients (specificity 88 %), and FP in 1 patient. For distant metastasis, PET/CT results were TP in 6 patients (sensitivity 100 %) and TN in 16 patients (specificity 100 %); there were no FN and FP cases. Conventional imaging was TP in 5 patients (sensitivity 83 %), FN in 1 patient, TN in 14 patients (specificity 88 %), and FP in 2 patients. The accuracy of PET/CT was higher than that of conventional imaging (91 vs 86 % for local recurrence and 100 vs 86 % for distant metastasis), but the differences were not statistically significant (p=0.1). Impact on treatment planning PET/CT findings resulted in a change in the treatment plan in 31.3 % (5/16) of staging cases and 13.8 % (4/29) of restaging cases (Table 3). There was no treatment change in surveillance cases. PET/CT had a high impact owing to a change from palliative to curative treatment in four patients (8.5 %). Figure 2 displays a sample case with a change in treatment
Table 3 Change in patient management by staging and restaging PET/CT Treatment planning by conventional imaging Staging Surgery Preoperative RT, surgery Chemotherapy Palliative chemoradiation Observation Restaging Surgery
Treatment planning by PET/CT
n
Reason for treatment change
Neoadjuvant therapy, surgery Chemoradiation, surgery Chemoradiation Surgery
1 1 1 1
Surgery
1
PET/CT upstaged from stage II to III PET/CT upstaged from stage II to III Additional bony metastatic lesion at left knee from PET/CT Lung and bone metastases on conventional imaging were shown to be benign by PET/CT Suspicion of malignant lesion from PET/CT
Surgery at different location
1
Chemotherapy Neoadjuvant therapy, surgery 3 Impact on treatment planning: staging 5/16 (31.3 %), restaging 4/29 (13.8 %) RT radiation therapy
False-positive tumor recurrence not shown on PET/CT. Additional avid lesion in abdomen seen on PET/CT PET/CT showed single lesion of disease
Eur J Nucl Med Mol Imaging Fig. 2 a, b CT showed primary tumor and lung metastasis (white arrow). c, d PET/CT showed primary tumor with benign lung lesion (yellow arrow). The treatment plan changed from systemic to local therapy. Pathology of the lung lesion was pulmonary hemangioma
from palliative to curative intent. Due to upstaging, PET/CT caused clinicians to give more adjuvant treatment in two patients and it detected additional metastatic disease in one
patient (6.4 %). PET/CT changed treatment from observation to surgery in one patient (2.1 %). Moreover, PET/CT changed the location of surgery in one patient (2.1 %). In this patient,
Fig. 3 Staging SUVmax (a) and restaging SUVmax (d) did not predict OS. Staging MTVand TLG (b, c) and restaging MTVand TLG (e, f) predicted OS
Eur J Nucl Med Mol Imaging Table 4 Prognostic factors for survival by univariate analysis (staging) Factor Age ≥40 years
