PET/MRI in Pancreatic and Periampullary Cancer: Correlating Diffusion-weighted Imaging, MR spectroscopy, and Glucose Metabolic Activity With Clinical Stage

Bang-Bin Chen¹, Yu-Wen Tien², Ming-Chu Chang³, Mei-Fang Cheng⁴, Yu-Ting Chang³, Chih-Horng Wu¹, Xin-Jia Chen¹, Ting-Chun Kuo², Shih-Hung Yang⁵, I-Lun Shih¹, Hong-Shiee Lai², and Tiffany Ting-Fang Shih¹

¹Medical Imaging and Radiology, National Taiwan University Medical School and Hospital, Taipei, Taiwan, ²Surgery, National Taiwan University Medical School and Hospital, Taipei, Taiwan, ³Internal Medicine, National Taiwan University Medical School and Hospital, Taipei, Taiwan, ⁴Nuclear Medicine, National Taiwan University Medical School and Hospital, Taipei, Taiwan, ⁵Oncology, National Taiwan University Medical School and Hospital, Taipei, Taiwan

Synopsis

We demonstrated that PET/MRI provides numerous useful imaging biomarkers for clinical staging and pathological grading in patients with pancreatic cancer or periampullary cancer. ADCmin was lower in tumors with N1 and an advanced TNM stage. Choline levels were higher in T4 and poorly differentiated tumors. Tumors with high glucose metabolic activity, as reflected by MTV and TLG, were at a more advanced T stage, exhibited lymph node and distant metastasis, and were at an advanced TNM stage. Moreover, compared with MTV or ADCmin alone, the MTV/ADCmin ratio demonstrated the highest predictive ability for determining the clinical TNM stage. Thus, integrated PET/MRI could provide complementary information on tumor characteristics, and these combined data could have stronger clinical or pathological implications than MRI or PET alone.

Purpose

To correlate the clinical TNM stage of pancreatic or periampullary cancer with the imaging biomarkers at diffusion-weighted imaging (DWI), magnetic resonance spectroscopy (MRS), and glucose metabolic activity derived from integrated positron emission tomography/magnetic resonance imaging (PET/MRI).

Methods

This prospective study was approved by the institutional review board and informed consent was obtained. Sixty consecutive patients (mean age, 62.7 ± 12.8 y; range, 24–85 y; 45 men, 15 women) with pancreatic (N=53) or periampullary cancer (N=7) underwent PET/MRI before treatment. The DWI (Figure 1, 2) was measured at the largest area of the tumor and the ADC map was calculated with a monoexponential function (b-values, 0, 600 and 1000 s/mm²). Single-voxel MRS data were acquired by point-resolved selective spectroscopy sequence with standard parameters (TR/TE, 1000/30 ms, FA, 90°) during free-breathing mode (Figure 3, 4). The imaging biomarkers were the minimal apparent diffusion coefficient (ADCmin), choline level from MRS, standard uptake values (SUVmax and SUVpeak), metabolic tumor volume (MTV), and total lesion glycolysis (TLG) of the tumors. The relationships between these biomarkers with pathologic findings and clinical TNM stage were evaluated using Pearson and Mann–Whitney U tests. The area under the receiver operating characteristic curve (AUROC) was used to evaluate accuracy.

Results

The MTV was higher in tumors with perineural invasion (P = .079); SUVmax (P = .079) and SUVpeak (P = .070) were both higher in tumors with lymphovascular invasion, but nonsignificant. A trend of increased choline (P = .073) and SUVpeak (P = .091) toward poor tumor differentiation was observed.

ADCmin correlated negatively with TLG (r = -0.284, P = .014) and was significantly lower in N1 (P = .027) and TNM Stage 3+ tumors (P = .043). Choline was significantly higher in T3+ tumors (P = .047). TLG was significantly higher in T3+ (P = .008), N1 (P = .033) and TNM Stage 3+ (P = .022) tumors. MTV was significantly higher in T3+, N1, M1, and TNM Stage 3+ tumors (all P < .05). Compared with ADCmin or MTV alone, the MTV/ADCmin ratio exhibited the highest AUROC for predicting T stage (T ≤ 3 vs T = 4, AUROC = 0.768), N stage (N = 0 vs N = 1, AUROC = 0.710), M stage (M = 0 vs M = 1, AUROC = 0.733), and advanced TNM stage (Stage ≤ 2 vs T ≥ 3, AUROC = 0.787) (Figure 5).

Discussion

Several studies have reported that ADC is useful in tissue characterization and prognosis prediction of pancreatic cancer. Tumors with low ADC values indicate shorter survival after resection [1-2]. Our results of a decreased ADCmin in tumors with N1 and an advanced TNM stage were consistent with the results of previous studies and could indicate poor prognosis in these patients. PET is an effective predictor of prognosis in patients with pancreatic cancer. MTV and TLG were significantly associated with baseline serum CA19-9 levels and also correlated with survival outcome after resection [3-4]. Elevated total choline derived with 1H MRS is a metabolic hallmark in multiple cancers, including pancreatic cancer [5].

Our results have important clinical implications. Highly aggressive tumors could lead to early metastasis after surgical resection, and short-term follow up or adjuvant chemotherapy might be necessary in such patients. Because both ADC and MTV are prognostic biomarkers in pancreatic or periampullary cancer, integrated PET/MRI is advantageous because it enables investigating these imaging biomarkers in a single examination.

Conclusion

Combining DWI, MRS, and PET data provided complementary information on tumor characteristics. MTV/ADCmin ratio had the highest accuracy for predicting clinical stage in patients with pancreatic or periampullary cancer.

Acknowledgements

The research is supported by National Taiwan University Medical School and Hospital and Ministry of Science and Technology, R.O.C.

References

1. Wang Y, Chen ZE, Nikolaidis P, McCarthy RJ, Merrick L, Sternick LA et al. Diffusion-weighted magnetic resonance imaging of pancreatic adenocarcinomas: association with histopathology and tumor grade. J Magn Reson Imaging. 2011;33:136-42.

2. Kurosawa J, Tawada K, Mikata R, Ishihara T, Tsuyuguchi T, Saito M et al. Prognostic relevance of apparent diffusion coefficient obtained by diffusion-weighted MRI in pancreatic cancer. J Magn Reson Imaging. 2015 May 6. doi: 10.1002/jmri.24939. [Epub ahead of print]

3. Lee JW, Kang CM, Choi HJ, Lee WJ, Song SY, Lee JH et al. Prognostic Value of Metabolic Tumor Volume and Total Lesion Glycolysis on Preoperative 18F-FDG PET/CT in Patients with Pancreatic Cancer. J Nucl Med. 2014;55:898-904.

4. Shi S, Ji S, Qin Y, Xu J, Zhang B, Xu W et al. Metabolic tumor burden is associated with major oncogenomic alterations and serum tumor markers in patients with resected pancreatic cancer. Cancer Lett. 2015;360:227-33.

5. Tesiram YA, Lerner M, Stewart C, Njoku C, Brackett DJ. Utility of nuclear magnetic resonance spectroscopy for pancreatic cancer studies. Pancreas. 2012;41:474-80.

Figures

Fig. 1 A 50 year-old male with pancreatic neck adenocarcinoma (T1N0M0, stage 1). DWI shows the a small hyperintense tumor (arrow) at pancreatic neck (b=600 s/mm²).

Fig. 2 A 50 year-old male with pancreatic neck adenocarcinoma (T1N0M0, stage 1). Hybrid PET and contrast-enhanced T1-weighted MR image clearly demonstrates high glucose metabolic activity of the tumor (arrow)

Fig. 3 A 60 year-old male with pancreatic body adenocarcinoma (T4N1M1, stage 4). Hybrid PET and contrast-enhanced T1-weighted MR image demonstrates high glucose metabolic activity of the tumor (arrows).

Fig. 4 A 60 year-old male with pancreatic body adenocarcinoma (T4N1M1, stage 4). MR spectroscopy shows high choline peak of the tumor at 3.2ppm (b)

Fig. 5 Receiver operating characteristic curves (AUROC) used to evaluate diagnostic performance of the ADCmin, MTV and MTV/ADCmin ratio for differentiating TNM stages in 60 pancreatic or periampullary cancers. MTV/ADCmin ratio exhibited the highest AUROC for predicting advanced TNM stage (Stage ≤ 2 vs T ≥ 3, AUROC = 0.787), compared with ADCmin and MTV respectively

Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)

0885