Pith. sign in

REVIEW 4 major objections 3 minor 88 references

Quantum Machine Learning for Predicting Anastomotic Leak: A Clinical Study

T0 review · 4 major / 3 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Simulated quantum neural networks trained on four clinical predictors generally outperform hyperparameter-tuned classical models at identifying patients at risk of anastomotic leak when the decision threshold is set for 83% sensitivity…

desk verdict A readable but methodologically flawed QML benchmark whose central comparative claim rests on test-set threshold fitting; worth a careful referee, not acceptance as is. read the letter →

arxiv 2506.01708 v2 pith:S2WPD6CH submitted 2025-06-02 quant-ph

Vojtěch Novák , Ivan Zelinka , Lenka Přibylová , Lubomír Martínek , Vladimír Benčurik This is my paper
classification quant-ph MSC 81P6868T0562P10
keywords anastomoticleakpredictionquantumneuralnetworksvariationalcircuitsZZFeatureMapEfficientSU2ansatzRealAmplitudesnoisesimulationclinicalrisk
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 tries to establish that small, noise-simulated quantum neural networks can serve as promising screening tools for anastomotic leak, a life-threatening complication after colorectal surgery, using a dataset of 200 patients. It builds four predictors (NoCoil, ACSP, DM, Smoking) and compares QNNs built from ZZFeatureMap encoding and two variational ansätze against grid-search-tuned classical models. At a fixed sensitivity of 83%, the QNNs generally score higher on accuracy, F1, PPV, NPV, and count R-squared, with the EfficientSU2-BFGS pair reaching a mean AUC of $0.7966 \pm 0.0237$. The authors also claim a direct link between optimizer convergence and downstream metrics, and an interpretability analysis suggesting QNNs capture non-linear feature interactions. They frame the results as preliminary, noting that the small sample size and single-center data require external validation before clinical use.

What carries the argument

The machinery is a hybrid variational quantum classifier: a ZZFeatureMap encodes the four clinical features into a 4-qubit state through single-qubit rotations and ZZ-interaction terms that create quadratic feature products, followed by a trainable ansatz (RealAmplitudes or EfficientSU2) whose parameters are optimized by BFGS, CMA-ES, COBYLA, SLSQP, or SPSA. The whole circuit is simulated under a depolarizing noise model with single-qubit gate error probability $p_{gate}=0.05$ and 1024 shots per circuit, approximating near-term hardware conditions. The argument runs on the coupling between ansatz expressiveness and optimizer convergence: configurations that reach lower final loss also show better AUC, calibration, and screening metrics, and the ZZFeatureMap's quadratic encoding is what the authors credit for the QNN's non-linear feature interactions.

What would settle it

Re-run the full pipeline with feature selection and threshold fitting performed only inside each training fold through nested cross-validation, then compare the QNNs against the same classical baselines at the fixed 83% sensitivity level; if the QNN advantage over the MLP and logistic regression disappears or reverses under this leak-free protocol, the central claim of general QNN superiority at fixed sensitivity is refuted.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is a trade-off between classification power and probability calibration. At a clinically motivated fixed sensitivity of 83%, simulated QNNs generally outperform classical models on threshold-dependent screening metrics, while classical models, particularly the multilayer perceptron, produce better-calibrated probabilities. The best QNN configuration, EfficientSU2 optimized by BFGS, achieved the highest mean AUC of $0.7966 \pm 0.0237$ and the lowest Brier score among QNNs ($0.1116 \pm 0.0029$); RealAmplitudes with CMA-ES led in average precision ($0.5041 \pm 0.1214$); and several QNN configurations reached negative predictive values of 95-96%. The paper argues that variational parameter optimization quality propagates directly into clinical metrics, and that the QNN's learned feature importance differs from logistic regression in ways that suggest non-linear, possibly entangled, feature interactions.

Load-bearing premise

The whole comparison rests on the assumption that outcome information never leaked into model construction: the paper itself concedes (Section X.D) that the sample is small with only 28 leak events, and the four predictors were chosen using the full dataset while the 83% sensitivity thresholds were fitted on the held-out fold, so any such leakage would make the reported AUC, NPV, and accuracy optimistic.

Editorial extensions

If this is right

  • If the 83%-sensitivity result holds, QNNs could serve as a first-pass screening tool for anastomotic leak, where high negative predictive value would reduce missed cases and guide postoperative monitoring.
  • The ansatz-optimizer choice matters more than the quantum-versus-classical label; no single configuration dominates all metrics, so deployment would require selecting a configuration to match the clinical goal.
  • Probability calibration is a separate axis from classification strength: classical models such as the MLP would remain preferred for risk scoring, while QNNs would be preferred for screening alerts that prioritize catching true positives.
  • The reported link between optimizer convergence and clinical metrics means that advances in variational optimization could translate directly into improved screening performance.
  • External validation on larger, multi-center cohorts is required before any clinical application, as the paper itself emphasizes.

Reading between the lines

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

  • A direct extension not tested in the paper would be to replace the ZZFeatureMap with a classical quadratic-feature baseline, such as logistic regression with explicit interaction terms, to isolate whether the quantum encoding's quadratic products are what drive the apparent gain.
  • The paper's fixed 83% sensitivity thresholds appear to have been selected using the held-out fold; a leak-free protocol that performs feature selection and threshold fitting entirely inside each training fold would test whether the QNN advantage over classical models survives a fully nested cross-validation.
  • A hardware test on a real noisy quantum device, which the paper lists as future work, would reveal whether the depolarizing error model with $p_{gate}=0.05$ is representative; if real-device noise erases the QNN advantage, the screening claim would be limited to simulation settings.
  • Since the dataset contains only 28 leak events, the reported NPV and AUC values have wide uncertainty intervals; a useful next analysis would report confidence intervals on the AUC difference between the best QNN and MLP rather than on each model separately.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 3 minor

Summary. The paper reports a benchmark of four-feature quantum neural network (QNN) classifiers against classical machine learning models for predicting anastomotic leak after colorectal surgery, using a single-center dataset of 200 patients (28 events). It claims that QNNs, simulated under a depolarizing noise model, generally outperform classical models at a fixed sensitivity of 83%, with the EfficientSU2-BFGS configuration achieving the highest AUC (0.7966 ± 0.0237), and it further discusses optimizer-convergence effects and perturbation-based feature importance. The stated clinical conclusion is that QNNs may serve as screening tools pending larger validation.

Significance. If the reported comparison were valid, the paper would provide a useful proof-of-concept for applying QNNs to a high-stakes, imbalanced clinical prediction task, with notable strengths: ten independent optimization runs per configuration, explicit noise modeling, and publicly available code and data access. However, the central comparative claim rests on an evaluation protocol that leaks test information at multiple stages, and the key summary table is internally inconsistent. As presented, the evidence does not support the QNN-over-classical claim, so the scientific contribution is currently not established.

major comments (4)
  1. [Section X.C and ROC appendix] The 'fixed sensitivity=83%' thresholds used for Table XIV are selected from the ROC curve of the held-out fold itself (Figures 24–44 state 'Fixed Sens (threshold=...)' for fold five), and the metrics in Table XIV are then reported on that same fold. This is test-set threshold fitting, which systematically inflates threshold-dependent metrics such as specificity, PPV, NPV, and accuracy. The paper's headline claim that 'QNNs generally outperform classical models across most metrics' relies on exactly these metrics; the calibration metrics (Brier Score, Log Loss) in the same table are comparable or favor the classical MLP. A valid protocol must select thresholds on training/validation data only, e.g., by nested cross-validation, and then evaluate once on the test fold.
  2. [Table XIV, Section X.C] The QNN rows of Table XIV are internally inconsistent. With a prevalence of 14% (28/200), accuracy at 83% sensitivity is determined by specificity: accuracy = 0.83×0.14 + specificity×0.86. For QNN-CMAES-RA (specificity 64%), the implied accuracy is about 67%, not the reported 84%; for QNN-BFGS-RA (specificity 35%), the implied accuracy is about 42%, not 83%. The reported accuracy values match the Count R2 column of the upper table, suggesting that the accuracy and the sensitivity/specificity columns are computed at different operating points. This makes Table XIV unreliable as a comparative table and undermines the conclusion drawn from it.
  3. [Sections VIII–IX and Appendix B] Feature selection is performed on the full dataset before any cross-validation: chi-square tests and AIC/BIC reduction (Sections VIII–IX) identify NoCoil, ACSP, DM, and Smoking using all 200 patients, and the same patients are then used for the cross-validated evaluations reported in Section X and Appendix B. This leaks outcome information into the feature-selection step and biases all subsequent AUC, accuracy, and calibration estimates. Correct methodology requires performing feature selection inside each training fold, or otherwise demonstrating that the selected features do not depend on the held-out outcomes.
  4. [Section X.B and Abstract] The abstract's headline result (EfficientSU2-BFGS with mean AUC 0.7966 ± 0.0237) is selected post hoc as the best among ten optimizer-ansatz configurations across multiple metrics. The same section notes that different configurations excel in different metrics and that 'no single configuration dominated across all metrics.' Reporting the maximum over configurations as the main result, without any multiple-comparison correction or pre-specified primary analysis, substantially inflates the apparent performance and makes the 'QNNs generally outperform' claim difficult to interpret.
minor comments (3)
  1. [Section VI, Table I] SPSA is classified as a 'Gradient-Free Method,' but SPSA is a stochastic gradient approximation technique that estimates gradients from two function evaluations per iteration; this classification should be corrected.
  2. [Figure 4] The caption 'Risk factors of Anastomotic leak' is uninformative; the text refers to 'the figure' with multiple panels, but the figure has no subplot labels or legend, making it difficult for the reader to map the visual claims to the variables NoCoil, ACSP, DM, and Smoking.
  3. [Section V.A] The text states that p_gate = 0.05 is a 'conservative upper bound' consistent with IBM Manila error rates, but typical single-qubit gate error rates on such devices are 0.1–0.5%, so 5% is roughly an order of magnitude higher; the justification for this choice should be clarified or revised.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the QNN-versus-classical benchmark is an empirical measurement with independent classical baselines, and the self-citations are background only.

full rationale

The paper's central results are measured, not derived: QNN AUC, average precision, calibration, and R2 values come from simulated noisy circuits evaluated under cross-validation, benchmarked against independently hyperparameter-tuned classical models. There is no equation in which a claimed output is defined in terms of that same output, and no fitted parameter is renamed as a prediction. The 'fixed sensitivity of 83%' protocol in Section X.C and the ROC appendix does select thresholds from the fifth-fold ROC and report metrics on that same fold; this is a test-set threshold-fitting and optimism concern, but it is not circular under the seven patterns because the sensitivity is an imposed operating point and the reported specificity, PPV, NPV, and accuracy are not equal by construction to any fitted input. Similarly, feature selection by chi-square and AIC on the full dataset (Sections VIII–IX) creates a leakage risk but no definitional circularity. The only self-references (refs [75] and [76]) support a background statement about CMA-ES robustness; the paper's own convergence experiments in Table XIII carry that claim, so those citations are not load-bearing. The paper explicitly flags the N=200 sample-size limitation in Section X.D and the Conclusion, which is a generalizability caveat rather than a circularity signal. Hence no significant circularity.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claim rests on the chosen noise level p_gate, the outcome-selected feature subset, and the test-set-tuned 83% threshold. These are free parameters chosen to produce the reported results. The main domain assumptions are representativeness of the single-center cohort, adequacy of the depolarizing noise model, and suitability of the QNN family.

free parameters (3)
  • p_gate = 0.05
    Chosen by hand as a conservative upper bound for single-qubit depolarizing error based on IBM Manila benchmarks; not measured for the simulated circuits and not varied.
  • feature_subset = NoCoil, ACSP, DM, Smoking
    Selected on the full dataset by univariate significance and AIC stepwise reduction; the selection used outcome information, making these effective free parameters chosen to fit the data.
  • sensitivity_target = 83%
    Hand-picked fixed sensitivity for comparing models; the thresholds that achieve it are fit on the test set.
assumptions (4)
  • domain assumption Patient cohort of 200 from a single hospital (2015-2016) is representative of the broader colorectal surgery population for AL prediction.
    The paper generalizes from one center and time window; no external cohort.
  • domain assumption Depolarizing noise with only single-qubit errors adequately approximates real noisy quantum hardware for benchmarking QNNs.
    Section V; no comparison with hardware, no two-qubit or readout errors.
  • domain assumption The ZZFeatureMap and chosen ansaetze form a suitable model family for binary clinical features.
    Standard QML choices; no theoretical or empirical justification specific to this problem.
  • standard math Standard statistical independence of patients and logistic regression assumptions.
    Used in statistical tests in Section VIII.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Quantum Machine Learning for Predicting Anastomotic Leak: A Clinical Study." pith.science (2026). https://pith.science/paper/S2WPD6CH

@misc{pith2026250601708,
  author       = {Pith},
  title        = {Pith review of: Quantum Machine Learning for Predicting Anastomotic Leak: A Clinical Study},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/S2WPD6CH}},
  note         = {Machine review of arXiv:2506.01708}
}
abstract

Anastomotic leak (AL) is a life-threatening complication following colorectal surgery, and its accurate prediction remains a significant clinical challenge. This study explores the potential of Quantum Neural Networks (QNNs) for AL prediction, presenting a rigorous benchmark against hyperparameter-tuned classical models including logistic regression, multilayer perceptrons, and boosting algorithms. Using a clinical dataset of 200 patients and four key predictors identified through statistical analysis, we evaluated QNNs with ZZFeatureMap encoding and EfficientSU2 and RealAmplitudes ans\"atze simulated under realistic hardware noise models. Our framework emphasizes robustness, with performance metrics averaged over 10 independent optimization runs using multiple algorithms. The EfficientSU2-BFGS combination achieved the highest mean AUC of $0.7966 \pm 0.0237$, while RealAmplitudes with CMA-ES excelled in Average Precision ($0.5041 \pm 0.1214$), critical for imbalanced medical datasets. We establish a direct link between optimizer convergence and model efficacy, where effective variational parameter optimization translates to improved classification metrics. Interpretability analysis suggests QNNs may capture complex, non-linear feature relationships not evident in classical linear models. This work highlights QNNs' potential as screening tools while underscoring the need for model selection based on specific clinical goals, pending validation on larger datasets.

Figures

Figures reproduced from arXiv: 2506.01708 by the authors.

Figure 2
Figure 2. The layered architecture allows for systematic construction of expressive quantum states while maintaining a relatively shallow cir￾cuit depth, which is crucial for maintaining co￾herence in noisy quantum environments. The Real Amplitudes ansatz with 4 qubits and 3 repetitions demonstrates favorable circuit characteristics with a depth of 11 and a total of 25 quantum gates, reflecting its design phi￾losophy of compu… view at source ↗
Figure 1
Figure 1. FIG. 1 [PITH_FULL_IMAGE:figures/full_fig_p010_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2 [PITH_FULL_IMAGE:figures/full_fig_p011_2.png] view at source ↗
Figures from the paper (29 more)
Figure 3
Figure 3. Figure 3: FIG. 3 [PITH_FULL_IMAGE:figures/full_fig_p012_3.png]
Figure 4
Figure 4. Figure 4: FIG. 4 [PITH_FULL_IMAGE:figures/full_fig_p017_4.png]
Figure 5
Figure 5. Figure 5: FIG. 5 [PITH_FULL_IMAGE:figures/full_fig_p021_5.png]
Figure 6
Figure 6. Figure 6: FIG. 6 [PITH_FULL_IMAGE:figures/full_fig_p022_6.png]
Figure 7
Figure 7. Figure 7: FIG. 7 [PITH_FULL_IMAGE:figures/full_fig_p023_7.png]
Figure 8
Figure 8. Figure 8: FIG. 8 [PITH_FULL_IMAGE:figures/full_fig_p023_8.png]
Figure 9
Figure 9. Figure 9: FIG. 9 [PITH_FULL_IMAGE:figures/full_fig_p024_9.png]
Figure 10
Figure 10. Figure 10: FIG. 10 [PITH_FULL_IMAGE:figures/full_fig_p030_10.png]
Figure 11
Figure 11. Figure 11: FIG. 11 [PITH_FULL_IMAGE:figures/full_fig_p030_11.png]
Figure 13
Figure 13. Figure 13: FIG. 13 [PITH_FULL_IMAGE:figures/full_fig_p032_13.png]
Figure 12
Figure 12. Figure 12: FIG. 12 [PITH_FULL_IMAGE:figures/full_fig_p032_12.png]
Figure 15
Figure 15. Figure 15: FIG. 15 [PITH_FULL_IMAGE:figures/full_fig_p033_15.png]
Figure 16
Figure 16. Figure 16: FIG. 16 [PITH_FULL_IMAGE:figures/full_fig_p033_16.png]
Figure 19
Figure 19. Figure 19: FIG. 19 [PITH_FULL_IMAGE:figures/full_fig_p034_19.png]
Figure 20
Figure 20. Figure 20: FIG. 20 [PITH_FULL_IMAGE:figures/full_fig_p034_20.png]
Figure 22
Figure 22. Figure 22: FIG. 22 [PITH_FULL_IMAGE:figures/full_fig_p036_22.png]
Figure 23
Figure 23. Figure 23: FIG. 23 [PITH_FULL_IMAGE:figures/full_fig_p037_23.png]
Figure 25
Figure 25. Figure 25: FIG. 25 [PITH_FULL_IMAGE:figures/full_fig_p041_25.png]
Figure 26
Figure 26. Figure 26: FIG. 26 [PITH_FULL_IMAGE:figures/full_fig_p041_26.png]
Figure 24
Figure 24. Figure 24: FIG. 24 [PITH_FULL_IMAGE:figures/full_fig_p041_24.png]
Figure 28
Figure 28. Figure 28: FIG. 28 [PITH_FULL_IMAGE:figures/full_fig_p042_28.png]
Figure 29
Figure 29. Figure 29: FIG. 29 [PITH_FULL_IMAGE:figures/full_fig_p042_29.png]
Figure 32
Figure 32. Figure 32: FIG. 32 [PITH_FULL_IMAGE:figures/full_fig_p043_32.png]
Figure 33
Figure 33. Figure 33: FIG. 33 [PITH_FULL_IMAGE:figures/full_fig_p043_33.png]
Figure 36
Figure 36. Figure 36: FIG. 36 [PITH_FULL_IMAGE:figures/full_fig_p044_36.png]
Figure 37
Figure 37. Figure 37: FIG. 37 [PITH_FULL_IMAGE:figures/full_fig_p044_37.png]
Figure 40
Figure 40. Figure 40: FIG. 40 [PITH_FULL_IMAGE:figures/full_fig_p045_40.png]
Figure 41
Figure 41. Figure 41: FIG. 41 [PITH_FULL_IMAGE:figures/full_fig_p045_41.png]
Figure 44
Figure 44. Figure 44: FIG. 44 [PITH_FULL_IMAGE:figures/full_fig_p046_44.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

88 extracted references · 65 canonical work pages

  1. [1]

    Regu- larization is applied to prevent overfitting

    Models and Hyperparameters • Logistic Regression : A linear model that predicts the probability of a binary outcome using a logistic function. Regu- larization is applied to prevent overfitting. – Possible Hyperparameters : Reg- ularization strength C ∈ {0.01, 0.1, 1, 10}, penalty type ∈{ L1, L2}. • K-Nearest Neighbors (KNN): A non- parametric model that ...

  2. [2]

    These parameters were selected based on the highest AUC score during the inner 3- fold cross-validation

    Optimal Hyperparameters The optimal hyperparameters for each classi- fier across the five folds are summarized in Ta- ble XV. These parameters were selected based on the highest AUC score during the inner 3- fold cross-validation. Appendix E: Quantum Measurement Operators The quantum neural network classification is performed through projective measuremen...

  3. [3]

    Carleo, Y

    G. Carleo, Y. Nomura, and M. Imada, Con- structing exact representations of quantum many-body systems with deep neural networks, Nature Communications 9, 1 (2018)

    2018

  4. [4]

    W. Liu, P. Gao, Y. Wang, W. Yu, and M. Zhang, A unitary weights based one- iteration quantum perceptron algorithm for non-ideal training sets, IEEE Access 7, 36854 (2019)

    2019

  5. [5]

    J. W. Z. Lau, K. H. Lim, H. Shrotriya, and L. C. Kwek, Nisq computing: where are we and where do we go?, AAPPS Bulletin 32, 27 (2022)

    2022

  6. [6]

    Biamonte, P

    J. Biamonte, P. Wittek, N. Pancotti, P. Reben- trost, N. Wiebe, and S. Lloyd, Quantum ma- chine learning, Nature 549, 195 (2017)

    2017

  7. [7]

    Ciliberto, M

    C. Ciliberto, M. Herbster, A. D. Ialongo, M. Pontil, A. Rocchetto, and S. Severini, Quan- tum machine learning: A classical perspective, Proc. Roy. Soc. A Math. Phys. Eng. Sci. 474, 10.1098/rspa.2017.0761 (2018)

    arXiv 2018

  8. [8]

    Schuld and N

    M. Schuld and N. Killoran, Quantum machine learning in feature hilbert spaces, Phys. Rev. Lett. 122, 40504 (2019)

    2019

Show all 88 references
  1. [9]

    J. D. Mart´ ın-Guerrero and L. Lamata, Quan- tum machine learning: A tutorial, Neurocom- puting 470, 457 (2022)

    2022

  2. [10]

    Wiebe, A

    N. Wiebe, A. Kapoor, and K. M. Svore, Quan- tum deep learning, Quantum Inf. Comput. 16, 541 (2016)

    2016

  3. [11]

    Somma, G

    R. Somma, G. Ortiz, E. Knill, and J. Guber- natis, Quantum simulations of physics prob- lems, Int. J. Quantum Inf. 1, 189 (2003)

    2003

  4. [12]

    Schuld, I

    M. Schuld, I. Sinayskiy, and F. Petruccione, An introduction to quantum machine learning, Contemp. Phys. 56, 172 (2015)

    2015

  5. [13]

    Sierra-Sosa, J

    D. Sierra-Sosa, J. Arcila-Moreno, B. Garcia- Zapirain, C. Castillo-Olea, and A. El- maghraby, Dementia prediction applying vari- ational quantum classifier, arXiv preprint arXiv:2007.08653 (2020)

    2020 arXiv

  6. [14]

    Schuld, Supervised quantum machine learn- ing models are kernel methods, arXiv preprint arXiv:2101.11020 (2021)

    M. Schuld, Supervised quantum machine learn- ing models are kernel methods, arXiv preprint arXiv:2101.11020 (2021)

    2021 arXiv

  7. [15]

    Havlicek, A

    V. Havlicek, A. D. Corcoles, K. Temme, A. W. Harrow, A. Kandala, J. M. Chow, and J. M. Gambetta, Supervised learning with quantum- enhanced feature spaces, Nature 567, 209 (2019)

    2019

  8. [16]

    Farhi and H

    E. Farhi and H. Neven, Classification with quantum neural networks on near term proces- sors, arXiv preprint arXiv:1802.06002 (2018)

    2018 arXiv

  9. [17]

    N. Moll, P. Barkoutsos, L. S. Bishop, J. M. Chow, A. Cross, D. J. Egger, and et al., Quan- tum optimization using variational algorithms on near-term quantum devices, Quantum Sci- ence and Technology 3, 30503 (2018)

    2018

  10. [18]

    Lloyd, M

    S. Lloyd, M. Mohseni, and P. Rebentrost, Quantum principal component analysis, Na- ture Physics 10, 631 (2014)

    2014

  11. [19]

    M. C. Caro, H.-Y. Huang, M. Cerezo, K. Sharma, A. Sornborger, L. Cincio, and P. J. Coles, Generalization in quantum machine learning from few training data, Nature Com- munications 13, 4919 (2022)

    2022

  12. [20]

    Kumar, G

    T. Kumar, G. Singh, and R. Kumar, Perfor- mance analysis of quantum classifier on bench- marking datasets, International Journal of Ad- vanced Computer Science and Applications10, 375 (2022)

    2022

  13. [21]

    Kharsa, A

    R. Kharsa, A. Bouridane, and A. Amira, Ad- vances in quantum machine learning and deep learning for image classification: A survey, Neurocomputing 552, 126506 (2023)

    2023

  14. [22]

    Jerbi, L

    S. Jerbi, L. J. Fiderer, H. P. Nautrup, J. M. K¨ ubler, H. J. Briegel, and V. Dun- jko, Qdataset: Quantum datasets for machine learning, arXiv preprint arXiv:2108.06661 (2021), arXiv:2108.06661 [quant-ph]

    2021 arXiv

  15. [23]

    Abohashima, M

    Z. Abohashima, M. Elhoseny, and E. H. Houssein, Classification with quantum ma- chine learning: A survey, arXiv preprint arXiv:2006.12270 (2020), arXiv:2006.12270 [quant-ph]

    2020 arXiv

  16. [24]

    Benˇ curik, M

    V. Benˇ curik, M. ˇSkrovina, L. Mart´ ınek, J. Bartoˇ s, M. Mach´ aˇ ckov´ a, M. Dosoudil, E. ˇStˇ ep´ anov´ a, L. Pˇ ribylov´ a, R. Briˇ s, and K. Vom´ aˇ ckov´ a, Intraoperative fluorescence an- giography and risk factors of anastomotic leak- age in mini-invasive low rectal ...

    2021

  17. [25]

    Blanco-Colino and E

    R. Blanco-Colino and E. Espin-Basany, Intra- operative use of icg fluorescence imaging to re- duce the risk of anastomotic leakage in col- orectal surgery: a systematic review and meta- analysis, Techniques in coloproctology 22, 15 (2018)

    2018

  18. [26]

    H. Kato, S. Munakata, K. Sakamoto, K. Sug- imoto, R. Yamamoto, S. Ueda, S. Tokuda, S. Sakuraba, T. Kushida, H. Orita, et al. , Im- 48 pact of left colonic artery preservation on anas- tomotic leakage in laparoscopic sigmoid resec- tion and anterior resection for sigmoid and re...

    2019

  19. [27]

    S. Zhao, L. Zhang, F. Gao, M. Wu, J. Zheng, L. Bai, F. Li, B. Liu, Z. Pan, J. Liu, et al. , Transanal drainage tube use for preventing anastomotic leakage after laparoscopic low an- terior resection in patients with rectal cancer: a randomized clinical trial, JAMA surgery156, ...

    2021

  20. [28]

    A. L. Axtell, C. Angeles, D. P. McCarthy, J. D. Maloney, G. E. Leverson, and M. M. De- Camp, Anastomotic leak after esophagectomy– analysis of the sts general thoracic surgery database, The Annals of Thoracic Surgery (2025)

    2025

  21. [29]

    Trastulli, G

    S. Trastulli, G. Munzi, J. Desiderio, R. Ciroc- chi, M. Rossi, and A. Parisi, Indocyanine green fluorescence angiography versus standard in- traoperative methods for prevention of anasto- motic leak in colorectal surgery: meta-analysis, British Journal of Surgery 108, 359 (2021)

    2021

  22. [30]

    J. He, M. He, J.-H. Tang, and X.-H. Wang, Anastomotic leak risk factors follow- ing colon cancer resection: a systematic re- view and meta-analysis, Langenbeck’s Archives of Surgery 408, 252 (2023)

    2023

  23. [31]

    S. Awad, A. I. Abd El-Rahman, A. Abbas, et al. , The assessment of perioperative risk factors of anastomotic leakage after intestinal surgeries: a prospective study, BMC Surgery 21, 29 (2021)

    2021

  24. [32]

    J. Lin, B. Zheng, S. Lin, et al. , The efficacy of intraoperative icg fluorescence angiography on anastomotic leak after resection for colorectal cancer: a meta-analysis, International Journal of Colorectal Disease 36, 27 (2021)

    2021

  25. [33]

    C. R. Balamurugan, R. Kodali, D. S. Hansalia, and V. Rengan, The impact of indocyanine green fluorescence angiography on anastomotic leak rates in colorectal surgery: a systematic review, Surgical Endoscopy 10.1007/s00464- 025-11547-1 (2025)

    2025 doi

  26. [34]

    M. L. Venn, R. L. Hooper, T. Pampiglione, et al. , Systematic review of preoperative and intraoperative colorectal anastomotic leak pre- diction scores (alps), BMJ Open 13, e073085 (2023)

    2023

  27. [35]

    Virtanen, R

    P. Virtanen, R. Gommers, T. E. Oliphant, M. Haberland, T. Reddy, D. Cournapeau, E. Burovski, P. Peterson, W. Weckesser, J. Bright, S. J. van der Walt, M. Brett, J. Wil- son, K. J. Millman, N. Mayorov, A. R. J. Nel- son, E. Jones, R. Kern, E. Larson, C. J. Carey, ˙I. Polat, Y. ...

    2020

  28. [36]

    Javadi-Abhari, M

    A. Javadi-Abhari, M. Treinish, K. Krsulich, C. J. Wood, J. Lishman, J. Gacon, S. Martiel, P. D. Nation, L. S. Bishop, A. W. Cross, B. R. Johnson, and J. M. Gambetta, Quantum com- puting with Qiskit (2024), arXiv:2405.08810 [quant-ph]

    2024 arXiv

  29. [37]

    Preskill, Quantum computing in the nisq era and beyond, Quantum 2, 79 (2018)

    J. Preskill, Quantum computing in the nisq era and beyond, Quantum 2, 79 (2018)

    2018

  30. [38]

    Farhi, J

    E. Farhi, J. Goldstone, and S. Gutmann, A quantum approximate optimization algorithm, arXiv preprint arXiv:1411.4028 (2014)

    2014 arXiv

  31. [39]

    Peruzzo, J

    A. Peruzzo, J. McClean, P. Shadbolt, M.-H. Yung, X.-Q. Zhou, P. J. Love, A. Aspuru- Guzik, and J. L. O’brien, A variational eigen- value solver on a photonic quantum processor, Nature communications 5, 4213 (2014)

    2014

  32. [40]

    Cerezo, A

    M. Cerezo, A. Arrasmith, R. Babbush, S. C. Benjamin, S. Endo, K. Fujii, J. R. McClean, K. Mitarai, X. Yuan, L. Cincio, et al. , Vari- ational quantum algorithms, Nature Reviews Physics 3, 625 (2021)

    2021

  33. [41]

    Schuld, I

    M. Schuld, I. Sinayskiy, and F. Petruccione, The quest for a quantum neural network, Quantum Information Processing 13, 2567 (2014)

    2014

  34. [42]

    Abbas, D

    A. Abbas, D. Sutter, C. Zoufal, A. Lucchi, A. Figalli, and S. Woerner, The power of quan- tum neural networks, Nature Computational Science 1, 403 (2021)

    2021

  35. [43]

    Benedetti, E

    M. Benedetti, E. Lloyd, S. Sack, and M. Fioren- tini, Parameterized quantum circuits as ma- chine learning models, Quantum Science and Technology 4, 043001 (2019)

    2019

  36. [44]

    M. C. Caro, H.-Y. Huang, M. Cerezo, K. Sharma, A. Sornborger, L. Cincio, and P. J. Coles, Generalization in quantum machine learning from few training data, Nature com- munications 13, 4919 (2022)

    2022

  37. [45]

    Larocca et al

    M. Larocca et al. , Group-invariant quantum machine learning, Physical Review A 106, 022402 (2022)

    2022

  38. [46]

    J. J. Meyer, M. Mularski, E. Gil-Fuster, A. A. 49 Mele, F. Arzani, A. Wilms, and J. Eisert, Exploiting symmetry in variational quantum machine learning, PRX quantum 4, 010328 (2023)

    2023

  39. [47]

    Y. Liu, S. Arunachalam, and K. Temme, A rig- orous and robust quantum speed-up in super- vised machine learning, Nat. Phys. 17, 1013 (2021)

    2021

  40. [48]

    F. J. Schreiber, J. Eisert, and J. J. Meyer, Clas- sical surrogates for quantum learning models, Physical Review Letters 131, 100803 (2023)

    2023

  41. [49]

    J¨ ager and R

    J. J¨ ager and R. V. Krems, Universal expres- siveness of variational quantum classifiers and quantum kernels for support vector machines, Nature Communications 14, 576 (2023)

    2023

  42. [50]

    Kandala, A

    A. Kandala, A. Mezzacapo, K. Temme, M. Takita, M. Brink, J. M. Chow, and J. M. Gambetta, Hardware-efficient variational quantum eigensolver for small molecules and quantum magnets, Nature 549, 242 (2017)

    2017

  43. [51]

    Hadfield, Z

    S. Hadfield, Z. Wang, B. O’Gorman, E. G. Ri- effel, D. Venturelli, and R. Biswas, From the quantum approximate optimization algorithm to a quantum alternating operator ansatz, Al- gorithms 12, 34 (2019)

    2019

  44. [52]

    K. Beer, D. Bondarenko, T. Farrelly, T. J. Osborne, R. Salzmann, D. Scheiermann, and R. Wolf, Training deep quantum neural net- works, Nature communications 11, 808 (2020)

    2020

  45. [53]

    Heimann, G

    D. Heimann, G. Sch¨ onhoff, and F. Kirch- ner, Learning capability of parametrized quan- tum circuits, arXiv preprint arXiv:2209.10345 (2022)

    2022 arXiv

  46. [54]

    Bud, History of biotechnology, e LS (2001)

    R. Bud, History of biotechnology, e LS (2001)

    2001

  47. [55]

    Van Beuzekom and A

    B. Van Beuzekom and A. Arundel, OECD biotechnology statistics 2006 (Citeseer, 2006)

    2006

  48. [56]

    Dabba, A

    A. Dabba, A. Tari, and S. Meftali, Hybridiza- tion of moth flame optimization algorithm and quantum computing for gene selection in mi- croarray data, Journal of Ambient Intelligence and Humanized Computing 12, 2731 (2021)

    2021

  49. [57]

    Sergioli, C

    G. Sergioli, C. Militello, L. Rundo, L. Minafra, F. Torrisi, G. Russo, K. L. Chow, and R. Giun- tini, A quantum-inspired classifier for clono- genic assay evaluations, Scientific Reports 11, 2830 (2021)

    2021

  50. [58]

    R. Y. Li, S. Gujja, S. R. Bajaj, O. E. Gamel, N. Cilfone, J. R. Gulcher, D. A. Lidar, and T. W. Chittenden, Quantum processor- inspired machine learning in the biomedical sci- ences, Patterns 2 (2021)

    2021

  51. [59]

    Sengupta and P

    K. Sengupta and P. R. Srivastava, Quantum algorithm for quicker clinical prognostic analy- sis: An application and experimental study us- ing ct scan images of covid-19 patients, BMC Medical Informatics and Decision Making 21, 1 (2021)

    2021

  52. [60]

    Acar and I

    E. Acar and I. Yilmaz, Covid-19 detection on ibm quantum computer with classical-quantum transfer learning, Turkish Journal of Electri- cal Engineering & Computer Sciences 29, 46 (2021)

    2021

  53. [61]

    Amin et al

    J. Amin et al. , Quantum machine learning ar- chitecture for covid-19 classification based on synthetic data generation using conditional ad- versarial neural network, Cognitive Computa- tion (2021)

    2021

  54. [62]

    El-shafeiy, A

    E. El-shafeiy, A. E. Hassanien, K. M. Sal- lam, and A. A. Abohany, Approach for training quantum neural network to predict severity of covid-19 in patients, Computational Materials Continua 66, 1745 (2021)

    2021

  55. [63]

    A. M. Iliyasu and C. Fatichah, A quantum hy- brid pso combined with fuzzy k-nn approach to feature selection and cell classification in cervi- cal cancer detection, Sensors 17, 2935 (2017)

    2017

  56. [64]

    Matic, M

    A. Matic, M. Monnet, J. M. Lorenz, B. Schachtner, and T. Messerer, Quantum- classical convolutional neural networks in radiological image classification (2022), arXiv:2204.12390 [quant-ph]

    2022 arXiv

  57. [65]

    Aishwarya, V

    S. Aishwarya, V. Abeer, B. B. Sathish, and K. Subramanya, Quantum computational tech- niques for prediction of cognitive state of hu- man mind from eeg signals, Journal of Quan- tum Computing 2, 157 (2020)

    2020

  58. [66]

    Gandhi, G

    V. Gandhi, G. Prasad, D. Coyle, L. Be- hera, and T. M. McGinnity, Quantum neu- ral network-based eeg filtering for a brain– computer interface, IEEE transactions on neu- ral networks and learning systems 25, 278 (2013)

    2013

  59. [67]

    Maheshwari, B

    D. Maheshwari, B. Garcia-Zapirain, and D. Sierra-Soso, Machine learning applied to di- abetes dataset using quantum versus classical computation, in 2020 IEEE international sym- posium on signal processing and information technology (ISSPIT) (IEEE, 2020) pp. 1–6

    2020

  60. [68]

    Gupta, H

    H. Gupta, H. Varshney, T. K. Sharma, N. Pachauri, and O. P. Verma, Compara- tive performance analysis of quantum machine learning with deep learning for diabetes pre- diction, Complex & Intelligent Systems 8, 3073 (2022)

    2022

  61. [69]

    Sierra-Sosa, J

    D. Sierra-Sosa, J. D. Arcila-Moreno, B. Garcia- Zapirain, and A. Elmaghraby, Diabetes type 2: 50 Poincare data preprocessing for quantum ma- chine learning., Computers, Materials & Con- tinua 67 (2021)

    2021

  62. [70]

    Maheshwari, U

    D. Maheshwari, U. Ullah, P. A. O. Marulanda, A. G.-O. Jurado, I. D. Gonzalez, J. M. O. Merodio, and B. Garcia-Zapirain, Quantum machine learning applied to electronic health- care records for ischemic heart disease classifi- cation, Hum.-Cent. Comput. Inf. Sci13 (2023)

    2023

  63. [71]

    S. Zhou, V. Giovannetti, S. Pirandola, E. M. Kessler, J. Preskill, and S. Lloyd, Achieving the heisenberg limit in quantum metrology us- ing quantum error correction, Nature Commu- nications 9, 78 (2018)

    2018

  64. [72]

    A. W. Cross, L. S. Bishop, S. Sheldon, P. D. Nation, and J. M. Gambetta, Validating quan- tum computers using randomized model cir- cuits, Physical Review A 100, 032328 (2019)

    2019

  65. [73]

    Vasques, H

    X. Vasques, H. Paik, and L. Cif, Applica- tion of quantum machine learning using quan- tum kernel algorithms on multiclass neuron m- type classification, Scientific Reports 13, 11541 (2023)

    2023

  66. [74]

    D. C. Liu and J. Nocedal, On the limited mem- ory bfgs method for large scale optimization, Mathematical programming 45, 503 (1989)

    1989

  67. [75]

    Hansen, S

    N. Hansen, S. D. M¨ uller, and P. Koumout- sakos, Reducing the time complexity of the derandomized evolution strategy with covari- ance matrix adaptation (cma-es), Evolutionary computation 11, 1 (2003)

    2003

  68. [76]

    Bonet-Monroig, H

    X. Bonet-Monroig, H. Wang, D. Vermetten, B. Senjean, C. Moussa, T. B¨ ack, V. Dunjko, and T. E. O’Brien, Performance comparison of optimization methods on variational quantum algorithms, Physical Review A 107, 032407 (2023)

    2023

  69. [77]

    Ill´ esov´ a, V

    S. Ill´ esov´ a, V. Nov´ ak, T. Bezdˇ ek, M. Beseda, and C. Possel, Numerical optimization strate- gies for the variational hamiltonian ansatz in noisy quantum environments, arXiv preprint arXiv:2505.22398 (2025)

    2025

  70. [78]

    Nov´ ak, I

    V. Nov´ ak, I. Zelinka, and V. Sn´ aˇ sel, Opti- mization strategies for variational quantum al- gorithms in noisy landscapes, arXiv preprint arXiv:2506.01715 (2025)

    2025

  71. [79]

    M. J. Powell, A direct search optimization method that models the objective and constraint functions by linear interpolation (Springer, 1994)

    1994

  72. [80]

    Kraft, A software package for sequential quadratic programming, Forschungsbericht- Deutsche Forschungs- und Versuchsanstalt fur Luft- und Raumfahrt (1988)

    D. Kraft, A software package for sequential quadratic programming, Forschungsbericht- Deutsche Forschungs- und Versuchsanstalt fur Luft- und Raumfahrt (1988)

    1988

  73. [81]

    J. C. Spall, Multivariate stochastic approxima- tion using a simultaneous perturbation gradi- ent approximation, IEEE transactions on au- tomatic control 37, 332 (1992)

    1992

  74. [82]

    N. R. Sproston and J. J. Ashworth, Role of c- reactive protein at sites of inflammation and in- fection, Frontiers in immunology 9, 754 (2018)

    2018

  75. [83]

    Akaike, Akaike’s information criterion, in International encyclopedia of statistical science (Springer, 2011) pp

    H. Akaike, Akaike’s information criterion, in International encyclopedia of statistical science (Springer, 2011) pp. 25–25

    2011

  76. [84]

    S. I. Vrieze, Model selection and psychological theory: a discussion of the differences between the akaike information criterion (aic) and the bayesian information criterion (bic)., Psycho- logical methods 17, 228 (2012)

    2012

  77. [85]

    McFadden, Regression-based specification tests for the multinomial logit model, Journal of econometrics 34, 63 (1987)

    D. McFadden, Regression-based specification tests for the multinomial logit model, Journal of econometrics 34, 63 (1987)

    1987

  78. [86]

    T. J. Smith and C. M. McKenna, A compar- ison of logistic regression pseudo r2 indices, Multiple Linear Regression Viewpoints 39, 17 (2013)

    2013

  79. [87]

    Ansin, An evaluation of the cox-snell resid- uals (2015)

    E. Ansin, An evaluation of the cox-snell resid- uals (2015)

    2015

  80. [88]

    J. Zhu, H. Zou, S. Rosset, T. Hastie, et al. , Multi-class adaboost, Statistics and its Inter- face 2, 349 (2009)

    2009

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.