Login to Access Video or Poster Abstract: MP52-05
Sources of Funding: none

Introduction

Near-infrared fluorescence (NIRF) technology in robotic surgery allows visualization of tissue fluorescence after giving intravenous indocyanine green (ICG). While this technology has been available for five years and has been adopted for perfusion assessment, it has not been widely adopted for achieving differential fluorescence (DF) of normal kidney versus tumors during robotic partial nephrectomy (RPN). Normal kidney fluoresces with NIRF due to bilitranslocase transport of ICG into proximal tubule cells while most renal tumors, including most renal cell carcinomas (RCC), should not fluoresce. This differential fluorescence of normal kidney and RCC allows better visualization of tumor edges and visual assessment of margins during and after tumor resection with fluorescence confirming a grossly negative margin. Most surgeons have abandoned using NIRF for DF citing inability to reliably achieve fluorescence of the kidney and not the tumor. While the dose of ICG is not critical when used to assess perfusion, improper ICG dosing will cause the tumor to fluoresce and make it indistinguishable from normal kidney. We developed an ICG dosing regimen providing reliable DF in most cases and report our success with NIRF.

Methods

RPN was performed with NIRF imaging in 253 tumors, the largest reported experience to date. Intraoperative assessment of tumor fluorescence was prospectively recorded at the time of surgery by the surgeon before pathologic assessment of tumors. The ICG dosing regimen included test doses beginning as low as 0.25cc to avoid overdosing and panfluorescence._x000D_

Results

Mean age was 58yrs (26-89) with mean body mass index of 32kg/m2 (18-63). Mean tumor size on imaging was 3.4cm (0.7-9.7) with mean R.E.N.A.L. nephrometry score of 7.3 (4-11), including 32 hilar tumors and 164 tumors >50% endophytic. Mean operative time was 170min. Among all 253 tumors, DF was successfully achieved in 217 (86%). Among 36 tumors that undesirably fluoresced, 8 were oncocytomas and 4 were chromophobe RCC, which are both known to express bilitranslocase, with 4 angiomyolipomas. Among 25 oncocytomas, 8 fluoresced such that DF was only successful in 68% of oncocytomas. Among 209 RCCs, 186 did not fluoresce (89%). Only 1 positive margin occurred (0.4%).

Conclusions

Robotic NIRF imaging was highly reliable (89%) in visualizing DF of RCC versus normal parenchyma but less so in oncocytomas as expected. Further study is needed to determine whether this contributed to the <1% positive margin rate.

Funding

none