Establishing the clinical utility of autofluorescence spectroscopy for parathyroid detection

Abstract

Background: The inability of surgeons to identify parathyroid glands accurately during cervical endocrine surgery hinders patients from achieving postoperative normocalcemia. An intrinsic, near-infrared fluorescence method was developed for real-time parathyroid identification with high accuracy. This study assesses the clinical utility of this approach.

Methods: Autofluorescence measurements were obtained from 137 patients (264 parathyroid glands) undergoing parathyroidectomy and/or thyroidectomy. Measurements were correlated to disease state, calcium levels, parathyroid hormone, vitamin D levels, age, sex, ethnicity, and body mass index. Statistical analysis identified which factors affect parathyroid detection.

Results: High parathyroid fluorescence was detected consistently and showed wide variability across patients. Near-infrared fluorescence was used to identify 256 of 264 (97%) of glands correctly. The technique showed high accuracy over a wide variety of disease states, although patients with secondary hyperparathyroidism demonstrated confounding results. Analysis revealed body mass index (P < .01), disease state (P < .01), vitamin D (P < .05), and calcium levels (P < .05) account greatly for variability in signal intensity. Age, sex, parathyroid hormone, and ethnicity had no effect.

Conclusion: This intrinsic fluorescence-based intraoperative technique can detect nearly all parathyroid glands accurately in real time. Its discrimination capacity is largely unlimited by patient variables, but several factors affect signal intensity. These results demonstrate potential clinical utility of optical guidance for parathyroid detection.

Copyright © 2016 Elsevier Inc. All rights reserved.

Figures

Figure 1

A) Typical near-infrared fluorescence spectra of parathyroid, thyroid, muscle, and fat in a single patient. B) Variability in parathyroid signal across 137 patients. The boxplot shows median, range, and interquartile range and outliers of normalized peak parathyroid signal. Black dots represent fluorescence intensity for each parathyroid measurement collected.

Figure 2

Average near infrared auto-fluorescence signal from parathyroid glands of patients with A) Normal (8.5 -10.5 mg/dL) and high (>10.5 mg/dL) serum calcium levels, B) Normal BMI (18.5 – 24.9 kg/m2) and high BMI (>25 kg/m2), C) Low (< 30 ng/dL) and high (30 – 80 ng/dL) total vitamin D levels, D) Benign thyroid disease-1, malignant thyroid disease-2, hyperthyroidism-3, hyperparathyroidism-4. Error bars indicate standard error of mean. Error bars indicate standard error of mean. Asterisks indicate significant differences (*p<0.05, **p<0.01, ***p<0.001)

Figure 3

A) Summary of glands removed from secondary hyperparathyroidism patients. Patient 4 had two glands removed with varying signal intensity. A) H&E stain of gland 1 of Patient 4 shows diffuse chief cell hyperplasia. B) H&E stain of gland 2 of Patient 4 shows water clear cell and oxyphil cell nodules amidst diffuse chief cell hyperplasia. (2x magnification with 10x inset).

Figure 4

Additional sources of variability. A) Intra-gland spatial heterogeneity contributes to variability in parathyroid and thyroid fluorescence signal. Each black dot corresponds to the fluorescence peak measured from a different location in a single gland. B) Probe contact pressure is inversely related to parathyroid fluorescence. In vitro measurements performed with controlled pressure from 0 – 60 N/cm2. In vivo measurements performed with surgeon’s assessment of “light”,“medium”, and “heavy” contact pressure.