UCSF

Background

Optimal treatment selection for localized renal tumors is challenging because of their variable biologic behavior and limitations in the preoperative assessment of tumor aggressiveness. The authors investigated the emerging hyperpolarized (HP) ¹³ C magnetic resonance imaging (MRI) technique to noninvasively assess tumor lactate production, which is strongly associated with tumor aggressiveness.

Methods

Eleven patients with renal tumors underwent HP ¹³ C pyruvate MRI before surgical resection. Tumor ¹³ C pyruvate and ¹³ C lactate images were acquired dynamically. Five patients underwent 2 scans on the same day to assess the intrapatient reproducibility of HP ¹³ C pyruvate MRI. Tumor metabolic data were compared with histopathology findings.

Results

Eight patients had tumors with a sufficient metabolite signal-to-noise ratio for analysis; an insufficient tumor signal-to-noise ratio was noted in 2 patients, likely caused by poor tumor perfusion and, in 1 patient, because of technical errors. Of the 8 patients, 3 had high-grade clear cell renal cell carcinoma (ccRCC), 3 had low-grade ccRCC, and 2 had chromophobe RCC. There was a trend toward a higher lactate-to-pyruvate ratio in high-grade ccRCCs compared with low-grade ccRCCs. Both chromophobe RCCs had relatively high lactate-to-pyruvate ratios. Good reproducibility was noted across the 5 patients who underwent 2 HP ¹³ C pyruvate MRI scans on the same day.

Conclusions

The current results demonstrate the feasibility of HP ¹³ C pyruvate MRI for investigating the metabolic phenotype of localized renal tumors. The initial data indicate good reproducibility of metabolite measurements. In addition, the metabolic data indicate a trend toward differentiating low-grade and high-grade ccRCCs, the most common subtype of renal cancer.

Lay summary

Renal tumors are frequently discovered incidentally because of the increased use of medical imaging, but it is challenging to identify which aggressive tumors should be treated. A new metabolic imaging technique was applied to noninvasively predict renal tumor aggressiveness. The imaging results were compared with tumor samples taken during surgery and showed a trend toward differentiating between low-grade and high-grade clear cell renal cell carcinomas, which are the most common type of renal cancers.