Geng Report: Multifunctional fiber-optic theranostic probe for closed-loop tumor photothermal therapy (Light: Science & Applications)

Academic fraud report · Geng Detector

Summary

This report assesses the 2026 Light: Science & Applications paper (DOI: 10.1038/s41377-026-02219-3) by Zesen Li, Zhuoran Li, Zhongyuan Cheng et al. for potential data integrity issues. Overall verdict: highly suspicious. Four concerns are flagged. (1) Figure 4c: stated temperatures of 22.7 °C at 0 mW and 102.9 °C at 267 mW are inconsistent with the authors' linear fit y = 27.601 + 0.275x, which predicts 27.6 °C at x = 0 and ~101.0 °C at x = 267; intermediate data points also yield non-linear slopes, suggesting the fitted equation does not match the plotted values. (2) The reported LnMOF excitation at 295 nm (deep UV) is unlikely to be transmitted effectively by a standard silica multimode fiber (400 μm core) over the in-vivo working length. (3) HCT116 xenograft inoculation at only 1×10⁵ cells in 20 μL Matrigel is far below the typical 1–5×10⁶ range, raising questions about the unusually rapid 100 mm³ growth timeline. (4) Figure 2j and Figure 3n share identical selectivity layouts, and n = 4/5 mice produced p < 0.0001, which is statistically unusual. Confidence is high for the mathematical inconsistency and low cell dose; lower for the fiber-physics and image concerns pending raw data.

Verdict

🔴 Highly suspicious. Multiple serious issues, including a verified mathematical inconsistency between the reported temperatures and the supplied linear fit equation in Figure 4c, plus method-level concerns about deep-UV excitation through a standard silica fiber and an unusually low xenograft cell dose.

Key findings

Figure 4c — Equation/data mismatch (severity: 🔴, verified). Reported endpoints are 22.7 °C at 0 mW and 102.9 °C at 267 mW. The authors' fit y = 27.601 + 0.275x yields 27.601 °C at x = 0 (not 22.7 °C) and ~101.0 °C at x = 267 (not 102.9 °C). Slopes between consecutive reported points (e.g., 0→44 mW: (43.1−22.7)/44 ≈ 0.463; 44→110 mW: (58.1−43.1)/66 ≈ 0.227) are not constant, contradicting a single linear relationship.
Methodology — Deep-UV excitation through standard silica fiber (severity: 🟠, under-substantiated). LnMOF is reported as excited at 295 nm. The Methods specify a standard multimode silica fiber with a 400 μm core. Standard high-OH silica strongly attenuates light below ~300 nm over short lengths; deep-UV or sapphire fibers would normally be required. No transmission-loss characterization is reported.
Methods — Low HCT116 inoculum (severity: 🟡, verified). Subcutaneous xenografts are established with only 1 × 10⁵ cells in 20 μL of Matrigel, well below the typical 1–5 × 10⁶ range, casting doubt on the feasibility of reaching ~100 mm³ within the days implied for therapy onset.
Figures 2j vs 3n and in-vivo statistics (severity: 🟠, partially verified). Figure 2j (pH selectivity) and Figure 3n (temperature selectivity) use the same interferent list and plotting template. With only n = 4–5 mice, multiple in-vivo comparisons reach p < 0.0001, an outcome that demands near-perfect inter-group separation and is uncommon in live-animal experiments.

Evidence highlights

Figure 4c endpoints: 0 mW → 22.7 °C; 267 mW → 102.9 °C.
Linear fit reported by authors: y = 27.601 + 0.275x.
Direct substitution contradictions: at x = 0, fit gives 27.601 °C (Δ = +4.901 °C vs reported); at x = 267, fit gives 101.026 °C (Δ = −1.874 °C vs reported).
Pairwise slopes from reported points: (0, 22.7) → (44, 43.1): ~0.463 °C/mW; (44, 43.1) → (110, 58.1): ~0.227 °C/mW; neither equals 0.275.
Fiber optics: reported excitation 295 nm through 400 μm-core multimode silica fiber; no UV-grade or sapphire-fiber specification or transmission-loss data provided.
Animal model: HCT116 at 1 × 10⁵ cells in 20 μL Matrigel; treatment initiated when tumors reportedly reached ~100 mm³.
Statistics: in-vivo groups with n = 4–5 returning p < 0.0001 across multiple endpoints.

Notes

DOI: 10.1038/s41377-026-02219-3 — preserved exactly as supplied.
All numeric values above are taken verbatim from the source report; no new measurements are inferred.
The deep-UV fiber concern is flagged as under-substantiated: confirming it requires either the original transmission spectra or vendor specifications for the specific fiber lot used.
Image-level reuse between Figures 2j and 3n cannot be confirmed from text alone and requires pixel-level comparison.
The author should be requested to provide (i) raw temperature-vs-power data underlying Figure 4c, (ii) fiber transmission/loss characterization at 295 nm, (iii) tumor growth curves with inoculation-to-treatment intervals, and (iv) raw in-vivo measurement spreadsheets to permit independent re-analysis.
This report is AI-assisted and intended for academic discussion only; formal determinations of misconduct require investigation by the journal and the authors' institution.