RapidET: a MEMS-based platform for label-free and rapid demarcation of tumors from normal breast biopsy tissues

Breast cancer is the most commonly diagnosed cancer in women and a major cause of mortality. Timely and accurate diagnosis, including intraoperative margin assessment, is critical but current methods such as frozen section analysis are time-consuming and require specialized pathology workflows. Tumor progression remodels the extracellular matrix and vasculature, altering electrical, thermal, and mechanical tissue properties. These biophysical changes can serve as label-free markers to objectively delineate tumor from adjacent normal tissue, particularly at surgical margins where differences are subtle. This study introduces RapidET, a MEMS-based system designed to rapidly quantify surface resistivity (ρs), bulk resistivity (ρb), and thermal conductivity (κ) of ex vivo breast biopsy tissues to distinguish tumors from adjacent normal tissues.

Prior MEMS technologies have demonstrated sensitive measurements at the single-cell and 2D cell culture scales (e.g., ECIS, microfluidics for CTC isolation and on-chip PCR/ICC), as well as microcantilever and PMUT-based approaches assessing mechanical and acoustic properties. While promising, many systems face challenges translating to bulk tissue characterization due to sample size constraints, complex operation (need for pumps, optics, high-speed imaging), and packaging robustness. Tissue-level applications have included 3D co-culture models (spheroids/organoids) and single-cell mechanical/electrical sensing, but compact, system-level platforms tailored for rapid probing of millimeter-scale ex vivo biopsy tissues remain limited. Measuring macroscale electrical and thermal properties (resistivity and conductivity) offers an objective, label-free avenue for tumor-normal delineation and assessment of tissue anisotropy due to matrix remodeling.

The RapidET system integrates silicon microchips with platinum microheaters, interdigitated electrodes (IDEs), and resistance temperature detectors (RTDs) via a slide-fit contact for rapid replacement. The platform includes mechatronic actuation and a GUI for control and data acquisition. Paired tumor and adjacent normal breast biopsy tissues from N=8 patients were tested in two routinely processed formats: deparaffinized FFPE sections and formalin-fixed fresh tissues. Prior to electrothermal measurements, samples were pathologist-diagnosed as normal or tumor.

Measurements: Tissue was heated using the on-chip microheater from 25 °C to 37 °C in 3 °C increments. Bulk resistivity (ρb) was measured across IDE pairs on two indenter subsystems (IS1 and IS2). Surface resistivity (ρs) was measured along the tissue surface using IDEs. Thermal conductivity (κ) was obtained using RTDs: the IS1 microheater served as the heat source, while RTDs (including the IS2 microheater used as an RTD) measured transmitted heat through the tissue. RTDs were calibrated to generate resistance–temperature profiles; the platinum microheater exhibited a temperature coefficient of resistance of 2.2×10^-3 °C^-1.

Statistical analysis: Paired two-tailed Student’s t-tests assessed within-group changes between 25 °C and 37 °C. Unpaired two-tailed Welch’s t-tests assessed tumor vs normal differences at each temperature. Fisher’s combined probability test and linear regression were used to evaluate the advantage of combining ρb, ρs, and κ for classification.

• Resistivity vs temperature: • Deparaffinized adjacent normal: mean ρb 148.42 ± 76.44 Ω·cm (25 °C) → 456.09 ± 194.63 Ω·cm (37 °C), 3.07×; not significant (P=0.103). • Deparaffinized tumor: mean ρb 411.25 ± 172.03 Ω·cm (25 °C) → 1980.87 ± 185 Ω·cm (37 °C), 4.82×; significant (P=0.0012). • Formalin-fixed adjacent normal: mean ρb 56.39 ± 10.4 Ω·cm (25 °C) → 195.70 ± 62.19 Ω·cm (37 °C), 3.47×; not significant (P=0.0957). • Formalin-fixed tumor: mean ρb 224.125 ± 61.72 Ω·cm (25 °C) → 991.4 ± 152.92 Ω·cm (37 °C), 4.42×; significant (P=0.014). • Surface resistivity showed similar trends and was higher in magnitude than bulk resistivity:
  • Deparaffinized adjacent normal: ρs 221.24 ± 74.71 Ω (25 °C) → 648.2 ± 419.44 Ω (37 °C), 2.93×; not significant (P=0.305).
  • Deparaffinized tumor: ρs 753.05 ± 292.05 Ω (25 °C) → 3131.88 ± 638.18 Ω (37 °C), 4.16×; significant (P=0.008).
  • Formalin-fixed adjacent normal: ρs 160.5 ± 24.25 Ω (25 °C) → 484.97 ± 159.8 Ω (37 °C); change not reported as significant.
  • Formalin-fixed tumor: ρs 507.08 ± 162.19 Ω (25 °C) → 2095.07 ± 116.83 Ω (37 °C); significant (fold 4.13, P=0.00744).
• Thermal conductivity (κ): tumor tissues exhibited lower κ than adjacent normal tissues for both sample types and temperatures. • Deparaffinized adjacent normal: κ 0.456 ± 0.023 W·m^-1·K^-1 (25 °C) vs 0.47 ± 0.018 W·m^-1·K^-1 (37 °C); no significant difference across temperatures (P=0.294). • Deparaffinized tumor: κ 0.207 ± 0.023 (25 °C) vs 0.255 ± 0.0255 (37 °C); significant across temperatures (P=0.00515). Between-group: normal > tumor at 25 °C (P=0.000602) and 37 °C (P=0.00149). • Formalin-fixed adjacent normal: κ 0.563 ± 0.028 (25 °C) vs 0.599 ± 0.022 (37 °C); significant across temperatures (P=0.0214). • Formalin-fixed tumor: κ 0.309 ± 0.02 (25 °C) vs 0.335 ± 0.0206 (37 °C); significant across temperatures (P=0.006). Between-group: normal > tumor at 25 °C (P=0.00103) and 37 °C (P=0.000321).
• Combining parameters: Fisher’s combined probability test and linear regression indicated that simultaneous use of ρb, ρs, and κ improves discrimination between tumor and adjacent normal tissues.

The RapidET system addresses the clinical need for rapid, label-free delineation of tumor versus adjacent normal breast tissue by quantifying electrical and thermal properties that reflect tumor-associated matrix remodeling and altered tissue architecture. Tumor tissues consistently exhibited higher surface and bulk resistivity and lower thermal conductivity compared with adjacent normal tissues, with statistically significant differences at 37 °C and at each temperature point for κ. Temperature-dependent increases in resistivity were more pronounced in tumors, supporting the sensitivity of electrothermal metrics to tumor phenotype. The ability to capture ρs, ρb, and κ on the same platform and the improved statistical significance when these metrics are combined suggest a robust, objective approach for tissue phenotyping that could aid margin assessment and streamline intraoperative or laboratory workflows.

This work presents RapidET, a compact MEMS-based electrothermal platform capable of rapid, label-free phenotyping of ex vivo breast biopsy tissues. By measuring surface resistivity, bulk resistivity, and thermal conductivity, the system differentiates tumor from adjacent normal tissue, with tumors showing higher resistivity and lower thermal conductivity across processing types and temperatures. The combined use of these metrics enhances statistical discrimination, highlighting the platform’s potential utility in pathology laboratories and operating rooms for faster, objective tissue assessment.

• Small cohort (N=8 patients) may limit generalizability and statistical power.
• Some within-group temperature-dependent changes (e.g., resistivity in adjacent normal tissues) were not statistically significant.
• Measurements were limited to 25–37 °C to avoid thermal instability of collagen above 37 °C, which may constrain exploration of broader thermal responses.
• Higher variability observed in some surface resistivity measurements (e.g., deparaffinized adjacent normal at 37 °C).
• Differences between deparaffinized and formalin-fixed preparations suggest processing-dependent variability that may affect absolute values and comparability.