Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1-10 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Per claim 1, for the limitation, “a processing device having a data processing module for extracting key feature data from the feature data of a patient to be predicted”, it is unclear whether “a patient to be predicted” is referring the extubation prediction for the patient or “extracting key feature data”, which can be a form of prediction. To expedite prosecution, Examiner interprets “a patient to be predicted” as ‘a patient subject to the extubation prediction’.
Per claims 1 and 7, for the last limitation, “obtaining a probability of extubation of
the patient to be predicted within 24 hours after the day of performing the extubation
prediction”, it is unclear if this limitation means the extubation prediction occurs within 24 hours or that the predicted extubation occurs within 24 hours. To expedite prosecution, Examiner assumes the former.
Claims 2-6 and 8-10 are rejected as being dependent upon a rejected base claim.
Claim Rejections - 35 USC § 102
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
Claims 1-10 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Explainable machine learning approach to predict extubation in critically ill ventilated patients: a retrospective study in central Taiwan to Pai et al. (hereinafter Pai, cited in IDS filed 6/21/2023).
Per claim 1, Pai discloses A system (pg. 10…computer system intrinsically used for executing code modules that implements the method/approached used, “All of the data and materials are provided in the manuscript and the supplemental data. The code has been put in public Github, and is available via https://github.com/GitTerrySu/Predict-extubation”) for establishing an extubation prediction using a machine learning model (pg. 1, Abstract…”we used an explainable machine learning (ML) approach to establish an extubation prediction model”), comprising:
a database for collecting feature data of patients (pg. 2, Methods…”The critical care database in this study was established through using data from the data warehouse at TCVGH…”), wherein each of the patients has a record (pg. 2, Methods…enrolled patients to hospital have data record on database, ”The critical care database in this study was established through using data from the data warehouse at TCVGH, a Taiwanese referral centre with approximately 1,500 beds and six ICUs in central Taiwan. Subjects who were admitted to ICUs between 2015 and 2019 were enrolled for analyses, and data of the first ICU admission was used among those with ICU admission more than one time”) of using a ventilatory assistance device during hospitalization (pg. 3, Results…patients were on mechanical ventilation, e.g., ventilatory assistance device, ”We enrolled 5,940 critically ill patients requiring mechanical ventilation for more than 48 h, and 65 features were used in the present study (Fig. 1)”), and the feature data comprise consciousness data (pg. 2, Methods…”We categorised the data into main clinical domains in accordance with the clinical workflow in critical care, and the four main clinical domains consisted of consciousness/awareness domain…the consciousness domain contained the Glasgow coma scale (GCS) as well as the Richmond Agitation Sedation Scale (RASS)”), input/output liquid data (pg. 2, Methods…”We categorised the data into main clinical domains in accordance with the clinical workflow in critical care, and the four main clinical domains consisted of…fluid balance domain…fluid balance domain included administered fluid, urine output as well as feeding amount…”), ventilatory function data (pg. 2, Methods…”We categorised the data into main clinical domains in accordance with the clinical workflow in critical care, and the four main clinical domains consisted of…ventilatory function domain…ventilatory parameter domain consisted of peak airway pressure (Ppeak), mean airway pressure (MAP), ventilator-day as well as respiratory rate…”) and physiological parameter data (pg. 2, Methods…”We categorised the data into main clinical domains in accordance with the clinical workflow in critical care, and the four main clinical domains consisted of…and physiological parameter domain…and physiology domain which was composed of heart rate”); and
a processing device having a data processing module (pg. 10…computer system intrinsically used for executing code modules using a processor, e.g., processing device, that implements the method/approached used, “All of the data and materials are provided in the manuscript and the supplemental data. The code has been put in public Github, and is available via https://github.com/GitTerrySu/Predict-extubation”) for extracting key feature data from the feature data of a patient to be predicted (pg. 2, Methods…limiting feature set to succinct features with feature elimination analysis is extracting key feature data, “We further applied recursive feature elimination for succinct features and used 20 features to establish the extubation prediction model (Supplemental, Fig. 3 for the results of recursive feature elimination analysis)”), wherein the key feature data comprise the physiological parameter data (pg. 5, fig. 3…Heart rate data, e.g., a physiological parameter data, is one of key features included "Heart rate (Day.-2)…Heart rate (Day.-3)" )), and the consciousness data (pg. 5, fig. 3…multiple key features associated with consciousness data, “GCS (Day.-2)…RASS (Day.-2)”), the input/output liquid data (pg. 5, fig. 3…”Urine (Day.-2)…Injection (Day.-2)”) and the ventilatory function data (pg. 5, fig. 3…”Ppeak (Day.-2…MAP (Day.-2)”) obtained 1 day or/and 2 days before a day of performing the extubation prediction (pg. 2, Methods…key feature data is data from two to three days prior to extubation, “Given that we aimed to predict weaning one day prior to extubation by using the two-day data (data of two and three days prior to extubation), the feature window and prediction window were hence 48 h and 24 h”); and an extubation prediction module (pg. 10…code modules are used to implement the extubation prediction functionality, including models used for analyzing key features, “All of the data and materials are provided in the manuscript and the supplemental data. The code has been put in public Github, and is available via https://github.com/GitTerrySu/Predict-extubation”) for analyzing the key feature data by using an extubation prediction model (pg. 2, Methods…XGBoost is one of the extubation prediction models employed, "We employed five machine learning (ML) models, including extreme gradient boosting (XGBoost)…"; pg. 1, Abstract…”We enrolled 5,940 patients and found the accuracy was comparable among XGBoost, LightGBM, CatBoost and RF, with the area under the receiver operating characteristic curve using XGBoost to predict extubation was 0.921”), and obtaining a probability of extubation of the patient to be predicted within 24 hours after the day of performing the extubation prediction (pg. 2, Methods…prediction window was 24 hours, "the feature window and prediction window were hence 48 h and 24 h, respectively"; pg. 9, fig. 6…LIME plots show "Extubation Probability: 0.81" and "0.19").
Per claim 2, Pai discloses claim 1, further disclosing the extubation prediction module (pg. 10…code modules are used to implement the extubation prediction functionality, “All of the data and materials are provided in the manuscript and the supplemental data. The code has been put in public Github, and is available via https://github.com/GitTerrySu/Predict-extubation”) analyzes the key feature data with the extubation prediction model, and then generates a correlated information between the key feature data and the probability of extubation (pg. 2, Methods…"In detail, the SHAP summary plot illustrated both the direction and strength of associations between key features and extubation probability and the partial dependence plot (PDP) further showed the marginal effect of the selected key features on the extubation prediction”).
Per claim 3, Pai discloses claim 2, further disclosing the processing device further comprises a visualization module (pg. 10…code modules are used to implement the extubation prediction functionality, “All of the data and materials are provided in the manuscript and the supplemental data. The code has been put in public Github, and is available via https://github.com/GitTerrySu/Predict-extubation”) for converting the correlated information between the key feature data and the probability of extubation into a visualization interface (figs. 4-6…visualization interfaces shown in Fig. 4 (SHAP summary plot), Fig. 5 (PDP plots), Fig. 6 (LIME and SHAP force plots); pgs. 2-3, Methods…"we showed extubation probability and used LIME and SHAP force plots for visualising the impact of key features on extubation").
Per claim 4, Pai discloses claim 1, further disclosing the extubation prediction model is selected from a group consisting of XGBoost (Extreme Gradient Boosting), CatBoost (categorical boosting), LightGBM (light gradient boosting machine) and random forest algorithm (RF) (pg. 2, Methods…”We employed five machine learning (ML) models, including extreme gradient boosting (XGBoost), categorical boosting (CatBoost), light gradient boosting machine (LightGBM), random forest (RF) and logistic regression (LR)…”).
Per claim 5, Pai discloses claim 1, further disclosing the processing device further comprises a model training module (pg. 10…code modules are used to implement the extubation prediction functionality including model training, “All of the data and materials are provided in the manuscript and the supplemental data. The code has been put in public Github, and is available via https://github.com/GitTerrySu/Predict-extubation”); the model training module trains a machine learning model with at least a part of the feature data of the patients (pg. 2, Methods…80/20 train/test split, ” We employed five machine learning (ML) models… and the ratio between training/testing was 80/20 in this study”), and verifies to generate the extubation prediction model (pg. 3, Methods…verification of the trained model by testing set, ”We determined the discrimination, accuracy and applicability of the models in the testing sets by the receiver operating characteristic (ROC) curve analysis, calibration curve as well as decision curve analysis [21, 22]. Python version 3.6 was applied in the present study”) and a key feature (pg. 2, Methods and fig. 3…feature importance calculated, “We further applied recursive feature elimination for succinct features and used 20 features to establish the extubation prediction model (Supplemental Fig. 3 for the results of recursive feature elimination analysis)”).
Per claim 6, Pai discloses claim 5, further disclosing the data processing module reads the feature data of the patients in the database (pg. 2, Methods… data extracted from data warehouse at TCVGH and features read for analysis, “The critical care database in this study was established through using data from the data warehouse at TCVGH…We categorised the data into main clinical domains in accordance with the clinical workflow in critical care, and the four main clinical domains”), and performs a data preprocessing procedure in order to delete data exceeding a preset reasonable range, or/and complement missing data (pg. 2, Methods…”With regards to data preprocessing, the physicians set the plausible range of each variable, and the missing data were imputed by the average value of each variable (Supplemental Table 1 for the plausible range and proportion of missing data of the top 20 variables with high feature importance)”).
Per claim 7, Pai discloses A method (pgs. 2-3, Methods) for establishing an extubation prediction using an machine learning model (pg. 1, Abstract…”we used an explainable machine learning (ML) approach to establish an extubation prediction model”), capable of predicting a possibility of extubation in real time (pg. 7…”right-aligned models can be used to continuously predict whether the target event will occur after the set time period, so-called real-time or continuous prediction models [25]. Therefore, the right-aligned design in the present study enables the proposed model to serve as an autonomous daily screen system to timely identify patients who were ready for breathing trial and to facilitate the weaning process through early recognition of the potential extubation one day earlier”), and comprising following steps:
step a: inputting feature data used for training (pg. 2, Methods…”The critical care database in this study was established through using data from the data warehouse at TCVGH…We employed five machine learning (ML) models, including extreme gradient boosting (XGBoost), categorical boosting (CatBoost), light gradient boosting machine (LightGBM), random forest (RF) and logistic regression (LR), and the ratio between training/testing was 80/20 in this study”), wherein the feature data are obtained from patients who use a ventilatory assistance device during hospitalization (pg. 3, Results…patients were on mechanical ventilation, e.g., ventilatory assistance device, ”We enrolled 5,940 critically ill patients requiring mechanical ventilation for more than 48 h, and 65 features were used in the present study (Fig. 1)”), and the feature data comprise consciousness data (pg. 2, Methods…”We categorised the data into main clinical domains in accordance with the clinical workflow in critical care, and the four main clinical domains consisted of consciousness/awareness domain…the consciousness domain contained the Glasgow coma scale (GCS) as well as the Richmond Agitation Sedation Scale (RASS)”), input/output liquid data (pg. 2, Methods…”We categorised the data into main clinical domains in accordance with the clinical workflow in critical care, and the four main clinical domains consisted of…fluid balance domain…fluid balance domain included administered fluid, urine output as well as feeding amount…”), ventilatory function data (pg. 2, Methods…”We categorised the data into main clinical domains in accordance with the clinical workflow in critical care, and the four main clinical domains consisted of…ventilatory function domain…ventilatory parameter domain consisted of peak airway pressure (Ppeak), mean airway pressure (MAP), ventilator-day as well as respiratory rate…”) and physiological parameter data (pg. 2, Methods…”We categorised the data into main clinical domains in accordance with the clinical workflow in critical care, and the four main clinical domains consisted of…and physiological parameter domain…and physiology domain which was composed of heart rate”);
step b: training a machine learning model by using at least a part of the feature data used for training (pg. 2, Methods…80/20 train/test split, ” We employed five machine learning (ML) models… and the ratio between training/testing was 80/20 in this study”) to generate an extubation prediction model (pg. 3, Methods…verification and generation of the trained model by testing set, ”We determined the discrimination, accuracy and applicability of the models in the testing sets by the receiver operating characteristic (ROC) curve analysis, calibration curve as well as decision curve analysis [21, 22]. Python version 3.6 was applied in the present study”) and a key feature (pg. 2, Methods and fig. 3…feature importance calculated, “We further applied recursive feature elimination for succinct features and used 20 features to establish the extubation prediction model (Supplemental Fig. 3 for the results of recursive feature elimination analysis)”), wherein the key feature comprises age (fig. 3…”AGE” being a key feature), a number of days of using a ventilator (fig. 3…”Ventilator-day” being a key feature), and GCS score (fig. 3…”GCS” being a key feature), urine volume (fig. 3…”Urine” being a key feature), injection volume (fig. 3…”Injection” being a key feature), nutrition amount (fig. 3…”Diet” being a key feature), RASS score (fig. 3…”RASS” being a key feature), PIP (fig. 3…”Ppeak” being a key feature), MAP (fig. 3…”MAP” being a key feature), ventilatory rate and heart rate of 1 day and 2 days before a day of performing the extubation prediction (fig. 3…”Respiratory rate (Day.-2)” is a key feature; fig. 3…”Heart rate (Day.-2)” is a key feature);
step c: inputting key feature data of a patient to be predicted (pg. 2, Methods…limiting feature set to succinct features with feature elimination analysis is extracting key feature data, “We further applied recursive feature elimination for succinct features and used 20 features to establish the extubation prediction model (Supplemental, Fig. 3 for the results of recursive feature elimination analysis)”), wherein the patient to be predicted is in a state of using a ventilatory assistance device (pg. 3, Results…patients were on mechanical ventilation, e.g., ventilatory assistance device, ”We enrolled 5,940 critically ill patients requiring mechanical ventilation for more than 48 h, and 65 features were used in the present study (Fig. 1)”); and
step d: analyzing the key feature data by using the extubation prediction model (pg. 2, Methods…XGBoost is one of the extubation prediction models employed, "We employed five machine learning (ML) models, including extreme gradient boosting (XGBoost)…"; pg. 1, Abstract…”We enrolled 5,940 patients and found the accuracy was comparable among XGBoost, LightGBM, CatBoost and RF, with the area under the receiver operating characteristic curve using XGBoost to predict extubation was 0.921”), and obtaining a probability of extubation of the patient to be predicted within 24 hours after a day of performing the extubation prediction (pg. 2, Methods…prediction window was 24 hours, "the feature window and prediction window were hence 48 h and 24 h, respectively"; pg. 9, fig. 6…LIME plots show "Extubation Probability: 0.81" and "0.19").
Per claim 8, Pai discloses claim 7, further disclosing the extubation prediction model is selected from a group consisting of XGBoost (Extreme Gradient Boosting), CatBoost (categorical boosting), LightGBM (light gradient boosting machine) and random forest algorithm (RF) (pg. 2, Methods…”We employed five machine learning (ML) models, including extreme gradient boosting (XGBoost), categorical boosting (CatBoost), light gradient boosting machine (LightGBM), random forest (RF) and logistic regression (LR)…”).
Per claim 9, Pai discloses claim 7, further disclosing a data preprocessing step for removing the feature data used for training that do not meet a standard, and/or complementing an insufficient part of the feature data used for training (pg. 2, Methods…”With regards to data preprocessing, the physicians set the plausible range of each variable, and the missing data were imputed by the average value of each variable (Supplemental Table 1 for the plausible range and proportion of missing data of the top 20 variables with high feature importance)”) by means of interpolation (imputation is a form of interpolation).
Per claim 10, Pai discloses claim 7, further disclosing obtaining a correlation between each of the key features and the probability of extubation (pg. 2, Methods…"In detail, the SHAP summary plot illustrated both the direction and strength of associations between key features and extubation probability and the partial dependence plot (PDP) further showed the marginal effect of the selected key features on the extubation prediction”), and converting the correlation into a visualization interface (figs. 4-6…visualization interfaces shown in Fig. 4 (SHAP summary plot), Fig. 5 (PDP plots), Fig. 6 (LIME and SHAP force plots); pgs. 2-3, Methods…"we showed extubation probability and used LIME and SHAP force plots for visualising the impact of key features on extubation").
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Patents and/or related publications are cited in the Notice of References Cited (Form PTO-892) attached to this action to further show the state of the art with respect to predicting mechanical ventilation extubation outcomes using machine learning.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ALAN CHEN whose telephone number is (571) 272-4143. The examiner can normally be reached M-F 10-7.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Kamran Afshar can be reached at (571) 272-7796. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/ALAN CHEN/Primary Examiner, Art Unit 2125