Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Autonomous Uncertainty Quantification for Computational Point-of-care Sensors

T0 review · 4 major / 5 minor · reviewed 2026-08-03 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A Monte Carlo dropout uncertainty check lets a point-of-care Lyme test flag its own unreliable predictions, lifting blinded sensitivity from 88.2% to 95.7%.

desk verdict First MCDO application to a computational POC sensor, with a real blind test, but the headline sensitivity gain is per-cartridge on 29 patients and needs patient-level verification. read the letter →

arxiv 2512.21335 v1 pith:4WVPTTOL submitted 2025-12-24 physics.med-ph cs.LGphysics.app-phphysics.bio-ph

Artem Goncharov , Rajesh Ghosh , Hyou-Arm Joung , Dino Di Carlo , Aydogan Ozcan This is my paper · ORCID
classification physics.med-phcs.LGphysics.app-phphysics.bio-ph
keywords uncertaintyquantificationMonteCarlodropoutpoint-of-carediagnosticsverticalflowassayLymediseaseneuralnetworksdiagnosticsensitivitycomputationalsensing
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper argues that neural-network-driven point-of-care sensors can police their own diagnostic outputs: by running each patient sample through many dropout-perturbed copies of the trained network and measuring how far the baseline prediction sits from the averaged perturbed predictions, the sensor can label its own prediction as trustworthy or not, with no need for ground-truth labels. On a paper-based vertical flow assay for Lyme disease, this autonomous reliability check improved diagnostic sensitivity on blinded patient samples from 88.2% to 95.7% by excluding high-uncertainty predictions and directing those samples to retesting. The central quantity is the figure of merit F = 1/|B0 - |, the reciprocal of the gap between the baseline score and the mean dropout score; low F flags unreliable predictions. If the approach generalizes, it gives computational POC sensors a built-in quality-assurance step that can reduce false negatives without changing the assay chemistry.

What carries the argument

The central object is the uncertainty figure of merit F = 1/|B0 − <MC>|, where B0 is the baseline model's sigmoid output for a sample and <MC> is the mean output of N Monte Carlo dropout models (same architecture, random dropout masks at 10% during inference). The reciprocal gap encodes predictive uncertainty: a small gap means the baseline prediction is stable under dropout perturbations and is trusted; a large gap (low F) means the prediction is unstable and is flagged 'Do not use'. The threshold Fth is a single scalar tuned on the validation set; the method requires only the trained network and its dropout layers, no ground-truth labels at inference.

What would settle it

On a new blinded cohort, compute F for every sample and also compute the baseline score alone; if excluding low-F samples removes no more errors than excluding samples with baseline scores nearest 0.5 (or than random exclusion with the same exclusion rate), the claim that F carries independent uncertainty information is falsified. Alternatively, if the optimal Fth on a second validation cohort differs widely from 8.5, the single-threshold generalization fails.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that the divergence between a baseline neural network's prediction score and the average prediction score of Monte Carlo dropout-perturbed versions of the same network is a usable, label-free proxy for diagnostic error in a computational point-of-care sensor. Defined as F = 1/|B0 - <MC>|, this figure of merit, with a single threshold (Fth = 8.5) tuned on a validation cohort, flagged a subset of predictions as unreliable. Excluding those flagged samples improved the sensitivity of the Lyme xVFA platform from 81.5% to 89.8% on validation data and, on an independent blinded cohort, from 88.2% to 95.7%, while also raising overall accuracy. The authors p

Load-bearing premise

The load-bearing premise is that the reciprocal gap between the baseline score and the average dropout-perturbed score, F = 1/|B0 - <MC>|, is a better indicator of a wrong prediction than the baseline score itself, and that a single threshold on F generalizes across patient cohorts.

Editorial extensions

If this is right

  • Samples flagged as 'Do not use' can be routed to repeat testing or gold-standard laboratory tests, reducing the clinical impact of false negatives.
  • The same MCDO-based quality-assurance step could be attached to other computational POC sensors (lateral flow, fluorescence, electrochemical) that use neural network inference.
  • Using as few as N=50 dropout models (0.66 s for 87 samples) matches the performance of N=1000, making the step practical for small, decentralized testing runs.
  • Low dropout rates (≤20%) during inference are preferable; larger rates destabilize the score distributions and degrade the filtering.
  • The approach adds a transparency layer for clinicians and regulators: each result carries an explicit reliability action, supporting trust in black-box diagnostic models.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A direct test of the FOM's validity would compare F against simpler alternatives (e.g., MC variance, entropy, or the baseline score's distance to the 0.5 threshold) on the same cohorts; the paper does not report such a comparison.
  • Because Fth is tuned on one validation set and applied to one blinded set, its portability to other populations, biobanks, or assay batches is an open question; the authors note that multi-center training data would be needed to set a unified threshold.
  • The sensitivity gain is achieved by excluding samples, not by reclassifying them; in a deployment, the excluded samples must be retested or referred, so the realized benefit depends on the follow-up pathway's own sensitivity and coverage.
  • The paper states that the same datasets were used for both training and validation due to limited sample size, which is a limitation worth monitoring; the FOM's calibration on the validation set could be optimistic for the blind test.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The manuscript presents an MCDO-based uncertainty quantification pipeline for a paper-based computational vertical flow assay (xVFA) for Lyme disease diagnosis. A baseline neural-network classifier (L0) is trained on 93 xVFA cartridges from 31 patients; the pipeline then runs N=1000 dropout-perturbed forward passes (10% dropout) and defines a figure of merit F = 1/|B0 - <MC>|, where B0 is the baseline score and <MC> is the average MCDO score. Samples with F below a threshold Fth=8.5 are labeled 'Do not use' and excluded from the reported diagnostic performance. Fth is optimized on the same 93-sample dataset used for training L0. On an independent blind cohort of 87 cartridges from 29 patients, excluding 5 samples (4 false negatives and 1 true positive) increases cartridge-level sensitivity from 88.2% to 95.7%. The authors also show that N=50 yields similar performance and that low dropout rates (<=20%) are preferable. The central claim is that this framework autonomously flags unreliable neural-network predictions without ground-truth labels, improving diagnostic sensitivity.

Significance. If the reported improvement is robust, the work offers a practical, low-overhead mechanism for adding a quality-assurance step to computational POC sensors—an important step toward clinical deployment. The blind-testing design is a genuine strength, as is the demonstration that the inference-time overhead is small (0.66 s for N=50). However, because the threshold Fth is selected using labels from the same samples that trained L0, and because the outcome is reported at the cartridge level for only 29 patients without confidence intervals or a patient-level analysis, the magnitude and clinical meaning of the sensitivity gain are not yet established. The concept is plausible and worth pursuing, but the evidence as presented is insufficient for publication in its current form.

major comments (4)
  1. [Methods, 'Baseline Lyme model architecture and training'] The Methods state: 'We utilized the same datasets for both training and validation due to the limited size of the training dataset.' This means that Fth=8.5 was tuned on the same 93 samples whose labels were used to train L0. The resulting validation-set improvement (81.5% to 89.8%) is therefore label-informed and likely optimistic. The blind test is independent, but the choice of Fth may still be overfit to idiosyncrasies of the training cohort. Please report the blind-test sensitivity for a range of Fth values (e.g., 5–12) to demonstrate that the improvement is not overly sensitive to the exact threshold, and ideally use a nested cross-validation or a separate validation cohort to select Fth.
  2. [Results, blind testing (Figure 5)] The headline sensitivity values (88.2% and 95.7%) are computed per cartridge, not per patient: 87 cartridges from only 29 patients. Replicates from the same patient are correlated, so treating cartridges as independent inflates the effective sample size. The paper does not report a patient-level contingency table, cluster-aware confidence intervals, or the number of patients contributing the excluded false-negative cartridges. Without this analysis, the 7.5-point improvement could be within sampling noise or driven by a single patient. Please provide patient-level sensitivity (e.g., majority vote or any-positive rule) and cluster-bootstrapped 95% CIs before and after filtering, and state how many patients had at least one cartridge flagged 'Do not use'.
  3. [Results, Eq. (1) and related text] The uncertainty figure of merit is F = 1/|B0 - <MC>|, and the authors note that 'the gap between these two sets of scores increases as the L0 output approaches the decision threshold of 0.5.' This raises the concern that F is largely a nonlinear transformation of the distance from the decision threshold. If so, a simpler abstention rule based on |B0 - 0.5| (a 'grey zone') might reproduce the same sensitivity gain. Please compare the proposed F-based filtering against such a baseline on the blind dataset—for example, exclude the same number of samples using a threshold on |B0 - 0.5| and report the resulting sensitivity. Without this comparison, the added value of the MCDO computation over the raw model score is not established.
  4. [Results/ Discussion, uncertainty calibration] The paper asserts that F is a valid, thresholdable proxy for prediction error, but it provides no calibration analysis and no comparison to alternative uncertainty metrics such as MCDO variance, predictive entropy, or MC interval width. The threshold Fth and the 10% dropout rate are free hyperparameters tuned on the validation set, and the SI shows that performance varies with dropout rate. Please add a calibration-style analysis (e.g., error rate versus F threshold) and a comparison with at least one alternative uncertainty metric, using a threshold selected only from the validation set and evaluated on the blind set. This would justify the specific definition of F and support the generalization of Fth across cohorts.
minor comments (5)
  1. [Introduction, first paragraph] Typo: 'disposal rapid diagnostic test' should be 'disposable rapid diagnostic test'.
  2. [Results, N=50 comparison] The claim that N=50 gives 'the same level of performance improvement' is based on a single threshold value and a single sensitivity point. Please report confidence intervals or a small bootstrap analysis for the sensitivity at N=50 to support the claim of comparability.
  3. [Results/Discussion, 'Do not use' handling] The reported post-filter sensitivity is conditional on samples passing the reliability check. The paper states that 'Do not use' samples can be sent for follow-up testing, but it does not report the proportion of samples flagged 'Do not use' (the inconclusive rate) or an intention-to-test analysis. Please report the inconclusive rate and discuss how the overall diagnostic sensitivity would change if follow-up testing resolves the excluded samples.
  4. [General] The supporting information figures S1–S5 are cited but not described in detail in the main text; ensure each supplementary figure is referenced in order and that their key results (especially dropout-rate sensitivity) are integrated into the main text's robustness discussion.
  5. [Data and code availability] The manuscript does not include a data and code availability statement. Given the role of the FOM definition and the importance of the threshold choice, providing the code and de-identified score distributions would aid reproducibility.

Circularity Check

1 steps flagged · score 2.0 of 10

Partial circularity only in the in-sample validation sensitivity gain; blind-test improvement is an independent holdout result.

  1. fitted input called prediction [Results and Discussion, 'Monte Carlo dropout (MCDO)-based autonomous uncertainty quantification for POC LD testing' (Fth optimization paragraph)]
    "The uncertainty threshold Fth was optimized on the validation dataset by selecting a single cutoff value that removed a substantial fraction of false predictions at the expense of missing a minimal number of correctly classified samples. By using Fth = 8.5 and excluding samples with lower F values, we filtered out 7 of the 11 misclassified samples in the validation dataset ... improving the validation test sensitivity of L0 from 81.5% to 89.8%"

    The validation-set sensitivity gain is an in-sample optimization artifact: Fth was chosen on the same 93 samples using ground-truth labels to remove false predictions, and those same samples were also used to train L0 ('We utilized the same datasets for both training and validation'). Thus the reported 81.5%→89.8% improvement is forced by the selection criterion, not predicted. This step is not load-bearing for the paper's central claim, because the abstract/conclusion highlight the independent blind-test result with Fth fixed before seeing those labels. Still, presenting the in-sample gain as 'validation test sensitivity' makes this a fitted-input-called-prediction step.

full rationale

The central empirical claim—blind-test sensitivity improving from 88.2% to 95.7% after MCDO filtering—is not circular. Fth=8.5 was set on the earlier dataset and then applied unchanged to 87 never-before-seen patient cartridges; the excluded samples on the blind test (4 of 6 false negatives, 1 true positive, 4 true negatives) are reported as outcomes, not as quantities used to fit the threshold. The FOM itself, F=1/|B0−<MC>|, is a fixed definition and was not tuned to blind-test labels. The only construction-reducing element is the validation-set sensitivity gain, which is an in-sample fit because the same 93 samples were used for training, Fth optimization, and the 'validation' evaluation. The paper discloses this overlap and does not rest its main contribution on it. Self-citations to prior xVFA/Lyme work are used as platform background, not as an unverified uniqueness argument, and no cited theorem is doing load-bearing work. Given the independent blind test, the circularity burden is low; score 2 reflects the one in-sample fitted 'prediction' while recognizing that the headline result survives as a genuine holdout evaluation.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The central claim rests on one fitted threshold Fth=8.5, a hand-selected 10% MCDO dropout rate, architecture hyperparameters selected on the same 93-sample data, and the ad hoc assumption that dropout-induced divergence from baseline marks unreliability. No new physical entities are proposed.

free parameters (3)
  • Fth = 8.5
    Uncertainty figure-of-merit threshold; optimized on the validation dataset to remove false predictions, then applied to blind testing. This is the key fitted constant of the reliability filter.
  • MCDO dropout rate = 10%
    Chosen among evaluated rates (1-90%); main results use 10%, with minor performance variations for ≤20% and worse performance at larger rates. A design choice that shapes the F distributions.
  • L0 architecture hyperparameters = 256/64 units, L2 λ=0.01, lr=1e-2, batch=8, training dropout=50%
    Selected via grid search on the same 93-sample dataset used for training and validation; these affect baseline L0 and therefore all downstream F values.
assumptions (5)
  • standard math Monte Carlo dropout approximates Bayesian model uncertainty (Gal & Ghahramani 2016).
    The method's validity builds on this prior result (cited [31]); dropout masks at inference are treated as samples from the posterior over network weights.
  • domain assumption The xVFA absorption signals x = 1 - s/b contain sufficient information for a neural network to infer Lyme disease status.
    Used as input to L0; assumes colorimetric spot intensities from the 25 spots map to IgM/IgG serostatus after background normalization, and that errors in L0 are detectable in this input space.
  • ad hoc to paper The divergence between baseline score and average MCDO score (F = 1/|B0-<MC>|) is a valid, thresholdable proxy for prediction error.
    Core assumption of the new reliability filter; no derivation or calibration evidence links F to error probability, and the reciprocal form is chosen heuristically.
  • domain assumption A single cutoff Fth=8.5 tuned on 31 patients from the Lyme Disease Biobank generalizes to new patients from the same biobank.
    The paper applies the validation-tuned threshold to the blind cohort; no multi-site or multi-biobank validation is provided, and the authors acknowledge generalization needs future expansion.
  • ad hoc to paper Training and validation on the same 93-sample dataset is acceptable for model selection.
    Stated in Methods as a necessity due to limited data; this means the validation metrics used to choose L0 and Fth are not independent of training, weakening the tuning step.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Autonomous Uncertainty Quantification for Computational Point-of-care Sensors." pith.science (2026). https://pith.science/paper/4WVPTTOL

@misc{pith2026251221335,
  author       = {Pith},
  title        = {Pith review of: Autonomous Uncertainty Quantification for Computational Point-of-care Sensors},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4WVPTTOL}},
  note         = {Machine review of arXiv:2512.21335}
}
read the original abstract

Computational point-of-care (POC) sensors enable rapid, low-cost, and accessible diagnostics in emergency, remote and resource-limited areas that lack access to centralized medical facilities. These systems can utilize neural network-based algorithms to accurately infer a diagnosis from the signals generated by rapid diagnostic tests or sensors. However, neural network-based diagnostic models are subject to hallucinations and can produce erroneous predictions, posing a risk of misdiagnosis and inaccurate clinical decisions. To address this challenge, here we present an autonomous uncertainty quantification technique developed for POC diagnostics. As our testbed, we used a paper-based, computational vertical flow assay (xVFA) platform developed for rapid POC diagnosis of Lyme disease, the most prevalent tick-borne disease globally. The xVFA platform integrates a disposable paper-based assay, a handheld optical reader and a neural network-based inference algorithm, providing rapid and cost-effective Lyme disease diagnostics in under 20 min using only 20 uL of patient serum. By incorporating a Monte Carlo dropout (MCDO)-based uncertainty quantification approach into the diagnostics pipeline, we identified and excluded erroneous predictions with high uncertainty, significantly improving the sensitivity and reliability of the xVFA in an autonomous manner, without access to the ground truth diagnostic information of patients. Blinded testing using new patient samples demonstrated an increase in diagnostic sensitivity from 88.2% to 95.7%, indicating the effectiveness of MCDO-based uncertainty quantification in enhancing the robustness of neural network-driven computational POC sensing systems.

Figures

Figures reproduced from arXiv: 2512.21335 by the authors.

Figure 1
Figure 1. Overview of the computational point-of-care sensing pipeline with uncertainty quantification. The pipeline outputs a reliability action for each sensor, classified as either “Trust” or “Do not use”. A “Trust” outcome indicates that the sensor measurement has passed reliability testing and that the corresponding diagnostic prediction can be used in the clinical decision-making process. In contrast, a “Do not use” out… view at source ↗
Figure 2
Figure 2. Monte Carlo dropout (MCDO)-based uncertainty quantification method for a classification neural network model. The baseline model score is compared with the distribution of scores generated by N MCDO models to perform uncertainty quantification in an autonomous manner, without access to ground truth patient information. Based on this comparison, each sample undergoes reliability assessment and is assigned one of the … view at source ↗
Figure 3
Figure 3. (a) Overview of the multiplexed vertical flow assay (xVFA) platform for Lyme disease diagnostics. (b) Monte Carlo dropout (MCDO)-based uncertainty quantification pipeline in the xVFA platform. For each xVFA sample, the score from the Lyme diagnostic model (L0) is compared with the score distribution from the MCDO Lyme models (Ln). Based on this comparison, an uncertainty figure of merit (F) is computed and evaluated… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: (a) Histogram, summarizing the Lyme model (L0) scores and the corresponding distributions of the MCDO-based Lyme model (Ln) outputs for 93 xVFA samples from the validation dataset, sorted from top to bottom by increasing L0 scores. (b) Uncertainty figure of merit (F) c…
Figure 5
Figure 5. Figure 5: Same as [PITH_FULL_IMAGE:figures/full_fig_p018_5.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Automated HER2 scoring with uncertainty quantification using lensfree holography and deep learning

    physics.med-ph 2026-01 conditional novelty 6.0 of 10

    Lensfree laser holography plus a five-network deep-learning ensemble scored HER2 status at 84.9% four-class / 94.8% binary accuracy on blinded patient cores, with Monte-Carlo-dropout uncertainty used to reject low-con...

Reference graph

Works this paper leans on

47 extracted references · cited by 1 Pith paper

  1. [1]

    & Mujawar, M

    Kaushik, A. & Mujawar, M. A. Point of care sensing devices: better care for everyone. Sensors 18, 4303 (2018)

    2018

  2. [2]

    & Lin, Z

    Zhang, W., Wang, R., Luo, F., Wang, P. & Lin, Z. Miniaturized electrochemical sensors and their point-of-care applications. Chin. Chem. Lett. 31, 589-600 (2020)

    2020

  3. [3]

    R., Goncharov, A., Eryilmaz, M., Ye, S., Palanisamy, B., et al

    Han, G. R., Goncharov, A., Eryilmaz, M., Ye, S., Palanisamy, B., et al. Machine learning in point-of- care testing: Innovations, challenges, and opportunities. Nat. Commun. 16, 3165 (2025)

    2025

  4. [4]

    S., Weckman, N

    Budd, J., Miller, B. S., Weckman, N. E., Cherkaoui, D., Huang, D., et al. Lateral flow test engineering and lessons learned from COVID-19. Nat. Rev. Bioeng. 1, 13-31 (2023)

    2023

  5. [5]

    Kakkar, S., Gupta, P., Yadav, S. P. S., Raj, D., Singh, G., et al. Lateral flow assays: Progress and evolution of recent trends in point-of-care applications. Mater. Today Bio 28, 101188 (2024)

    2024

  6. [6]

    M., Memeu, D

    Kinyua, D. M., Memeu, D. M., Mugo Mwenda, C. N., Ventura, B. D., & Velotta, R. Advancements and applications of lateral flow assays (LFAs): A comprehensive review. Sensors 25, 5414 (2025)

    2025

  7. [7]

    G., Lee, Y

    Jung, Y ., Kim, S., Kim, M. G., Lee, Y . E., Shin, M. G., & Yang, S. One-step detection of vancomycin in whole blood using the lateral flow immunoassay. Biosensors 14, 129 (2024)

    2024

  8. [8]

    R., & Kim, M

    Han, G. R., & Kim, M. G. Highly sensitive chemiluminescence-based lateral flow immunoassay for cardiac troponin I detection in human serum. Sensors 20, 2593 (2020)

    2020

Show all 47 references
  1. [9]

    T., Jin, E., & Lee, M

    Kim, H. T., Jin, E., & Lee, M. H. Portable chemiluminescence-based lateral flow assay platform for the detection of cortisol in human serum. Biosensors 11, 191 (2021)

    2021

  2. [10]

    Jung, C., & Kim, M. G. Direct use of a saliva-collected cotton swab in lateral flow immunoassay for the detection of cotinine. Biosensors 12, 214 (2022)

    2022

  3. [11]

    A., Kaushik, P., Jain, R

    Patil, A. A., Kaushik, P., Jain, R. D., & Dandekar, P. P. Assessment of urinary biomarkers for infectious diseases using lateral flow assays: a comprehensive overview. ACS Infect. Dis. 9, 9-22 (2022)

    2022

  4. [12]

    M., & Ozcan, A

    Ballard, Z., Brown, C., Madni, A. M., & Ozcan, A. Machine learning and computation-enabled intelligent sensor design. Nat. Mach. Intell. 3, 556-565 (2021)

    2021

  5. [13]

    R., Ghosh, R., Palanisamy, B., Di Carlo, D., Ozcan, A., & Park, S

    Jeon, J., Choi, H., Han, G. R., Ghosh, R., Palanisamy, B., Di Carlo, D., Ozcan, A., & Park, S. Paper- based Vertical Flow Assays for in Vitro Diagnostics and Environmental Monitoring. ACS sensors 10, 3317-3339 (2025)

    2025

  6. [14]

    A., Ghosh, R., Han, G

    Goncharov, A., Joung, H. A., Ghosh, R., Han, G. R., Ballard, Z. S., et al. Deep Learning‐Enabled Multiplexed Point‐of‐Care Sensor using a Paper‐Based Fluorescence Vertical Flow Assay. Small 19, 2300617 (2023)

    2023

  7. [15]

    A., Ballard, Z

    Joung, H. A., Ballard, Z. S., Wu, J., Tseng, D. K., Teshome, H., et al. Point-of-care serodiagnostic test for early-stage Lyme disease using a multiplexed paper-based immunoassay and machine learning. ACS nano 14, 229-240 (2019)

    2019

  8. [16]

    A paper-based assay for the colorimetric detection of SARS-CoV-2 variants at single-nucleotide resolution

    Zhang, T., Deng, R., Wang, Y ., Wu, C., Zhang, K., et al. A paper-based assay for the colorimetric detection of SARS-CoV-2 variants at single-nucleotide resolution. Nat. Biomed. Eng. 6, 957-967 (2022)

    2022

  9. [17]

    R., Joung, H

    Eryilmaz, M., Goncharov, A., Han, G. R., Joung, H. A., Ballard, Z. S., et al. A paper-based multiplexed serological test to monitor immunity against SARS-COV-2 Using machine learning. ACS nano 18, 16819-16831 (2024); 12

    2024

  10. [18]

    S., Joung, H

    Ballard, Z. S., Joung, H. A., Goncharov, A., Liang, J., Nugroho, K., Di Carlo, D., Garner, O. B., & Ozcan, A. Deep learning-enabled point-of-care sensing using multiplexed paper-based sensors. NPJ digit. Med. 3, 66 (2020)

    2020

  11. [19]

    A., Goncharov, A., Palanisamy, B., Ngo, K., et al

    Ghosh, R., Joung, H. A., Goncharov, A., Palanisamy, B., Ngo, K., et al. Rapid single-tier serodiagnosis of Lyme disease. Nat. commun. 15, 7124 (2024)

    2024

  12. [20]

    R., Goncharov, A., Eryilmaz, M., Joung, H

    Han, G. R., Goncharov, A., Eryilmaz, M., Joung, H. A., Ghosh, R., et al. Deep learning-enhanced paper-based vertical flow assay for high-sensitivity troponin detection using nanoparticle amplification. ACS nano 18, 27933-27948 (2024)

    2024

  13. [21]

    R., Goncharov, A., Eryilmaz, M., Ye, S., Joung, H

    Han, G. R., Goncharov, A., Eryilmaz, M., Ye, S., Joung, H. A., Deep Learning‐Enhanced Chemiluminescence Vertical Flow Assay for High‐Sensitivity Cardiac Troponin I Testing. Small 21, 2411585 (2025)

    2025

  14. [22]

    hallucinations

    Hatem, R., Simmons, B., & Thornton, J. E. A call to address AI “hallucinations” and how healthcare professionals can mitigate their risks. Cureus 15, (2023)

    2023

  15. [23]

    Zhang, J., & Zhang, Z. M. Ethics and governance of trustworthy medical artificial intelligence. BMC Med. Inform. Decis. Mak. 23, 7 (2023)

    2023

  16. [24]

    Wadden, J. J. Defining the undefinable: the black box problem in healthcare artificial intelligence. J. Med. Ethics 48, 764-768 (2022)

    2022

  17. [25]

    Trustworthy clinical AI solutions: A unified review of uncertainty quantification in Deep Learning models for medical image analysis

    Lambert, B., Forbes, F., Doyle, S., Dehaene, H., & Dojat, M. Trustworthy clinical AI solutions: A unified review of uncertainty quantification in Deep Learning models for medical image analysis. Artif. Intell. Med. 150, 102830 (2024)

    2024

  18. [26]

    Abdar, M., Khosravi, A., Islam, S. M. S., Acharya, U. R., & Vasilakos, A. V . The need for quantification of uncertainty in artificial intelligence for clinical data analysis: increasing the level of trust in the decision-making process. IEEE SMC 8, 28-40 (2022)

    2022

  19. [27]

    Aleatoric and epistemic uncertainty in machine learning: An introduction to concepts and methods

    Hüllermeier, E., & Waegeman, W. Aleatoric and epistemic uncertainty in machine learning: An introduction to concepts and methods. Mach. Learn. 110, 457-506 (2021)

    2021

  20. [28]

    Why do errors arise in artificial intelligence diagnostic tools in histopathology and how can we minimize them?

    Evans, H., & Snead, D. Why do errors arise in artificial intelligence diagnostic tools in histopathology and how can we minimize them?. Histopathol. 84, 279-287 (2024)

    2024

  21. [29]

    A., Cellini, J., Charpignon, M

    Celi, L. A., Cellini, J., Charpignon, M. L., Dee, E. C., Dernoncourt, F. Sources of bias in artificial intelligence that perpetuate healthcare disparities—A global review. PLOS digit. health 1, e0000022 (2022)

    2022

  22. [30]

    D., & Chang, D

    Flynn, C. D., & Chang, D. Artificial intelligence in point-of-care biosensing: challenges and opportunities. Diagnostics 14, 1100 (2024)

    2024

  23. [31]

    Dropout as a bayesian approximation: Representing model uncertainty in deep learning

    Gal, Y ., & Ghahramani, Z. Dropout as a bayesian approximation: Representing model uncertainty in deep learning. PMLR, 1050-1059 (2016)

    2016

  24. [32]

    H., Kim, B

    Kwon, Y ., Won, J. H., Kim, B. J., & Paik, M. C. Uncertainty quantification using Bayesian neural networks in classification: Application to biomedical image segmentation. Comput. Stat. Data Anal. 142, 106816 (2020)

    2020

  25. [33]

    Y ., et al

    Avci, M. Y ., et al. Improving accuracy and uncertainty quantification of deep learning based quantitative MRI using Monte Carlo dropout. arXiv Preprint at (2021)

    2021

  26. [34]

    H., et al

    Ahn, S. H., et al. Uncertainty Quantification in Automated Detection of Vertebral Metastasis Using Ensemble Monte Carlo Dropout. JIIM, 1-16 (2024)

    2024

  27. [35]

    I., et al

    Orlando, J. I., et al. U2-net: A bayesian u-net model with epistemic uncertainty feedback for photoreceptor layer segmentation in pathological oct scans. ISBI, 1441-1445 (2019)

    2019

  28. [36]

    Steere, A. C. Lyme disease. N. Engl. J. Med. 345, 115–125 (2001)

    2001

  29. [37]

    R., et al

    Bobe, J. R., et al. Recent progress in Lyme disease and remaining challenges. Front. Med. 8, 666554 (2021)

    2021

  30. [38]

    P., Gray, J., & Strle, F

    Stanek, G., Wormser, G. P., Gray, J., & Strle, F. Lyme borreliosis. The Lancet 379, 461-473 (2012); 13

    2012

  31. [39]

    Vital signs: trends in reported vectorborne disease cases—United States and Territories, 2004–2016

    Rosenberg, R. Vital signs: trends in reported vectorborne disease cases—United States and Territories, 2004–2016. MMWR 67, (2018)

    2004

  32. [40]

    & Schriefer, M

    Moore, A., Nelson, C., Molins, C., Mead, P. & Schriefer, M. Current guidelines, common clinical pitfalls, and future directions for laboratory diagnosis of Lyme disease, United States. Emerg. Infect. Dis. 22, 1169–1177 (2016)

    2016

  33. [41]

    Branda, J. A. et al. Advances in serodiagnostic testing for Lyme disease are at hand. Clin. Infect. Dis. 66, 1133–1139 (2018)

    2018

  34. [42]

    Joung, H.-A. et al. Point-of-care serodiagnostic test for early-stage Lyme disease using a multiplexed paper-based immunoassay and machine learning. ACS Nano 14, 229–240 (2020)

    2020

  35. [43]

    M., Petzke, M

    Signorino, G., Arnaboldi, P. M., Petzke, M. M. & Dattwyler, R. J. Identification of OppA2 linear epitopes as serodiagnostic markers for Lyme disease. Clin. Vaccin. Immunol. 21, 704–711 (2014)

    2014

  36. [44]

    Toumanios, C., Prisco, L., Dattwyler, R. J. & Arnaboldi, P. M. Linear B cell epitopes derived from the multifunctional surface lipoprotein BBK32 as targets for the serodiagnosis of Lyme disease. mSphere 4, e00111–e00119 (2019)

    2019

  37. [45]

    M., Katseff, A

    Arnaboldi, P. M., Katseff, A. S., Sambir, M. & Dattwyler, R. J. Linear peptide epitopes derived from ErpP, p35, and FlaB in the serodiagnosis of Lyme disease. Pathogens 11, 944 (2022)

    2022

  38. [46]

    Wormser, G. P. et al. The clinical assessment, treatment, and prevention of Lyme disease, human granulocytic anaplasmosis, and Babesiosis: clinical practice guidelines by the Infectious Diseases Society of America. Clin. Infect. Dis. 43, 1089–1134 (2006)

    2006

  39. [47]

    Trust” or “Do not use

    Proposed Regulatory Framework for Modifications to Artificial Intelligence/Machine Learning (AI/ML)-Based Software as a Medical Device (SaMD), U.S. Food & Drug Administration webpage, https://www.fda.gov/files/medical%20devices/published/US-FDA-Artificial-Intelligence-and- Mac...

    2019

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.