Enhanced patient-based real-time quality control using the graph-based anomaly detection.

OBJECTIVES: Patient-based real-time quality control (PBRTQC) is an alternative tool for laboratories that has gained increasing attention. Despite the progress made by using various algorithms, the problems of data volume imbalance between in-control and out-of-control results, as well as the issue of variation remain challenges. We propose a novel integrated framework using anomaly detection and graph neural network, combining clinical variables and statistical algorithms, to improve the error detection performance of patient-based quality control. METHODS: The testing results of three representative analytes (sodium, potassium, and calcium) and eight independent variables of patients (test date, time, gender, age, department, patient type, and reference interval limits) were collected. Graph-based anomaly detection network was modeled and used to generate control limits. Proportional and random errors were simulated for performance evaluation. Five mainstream PBRTQC statistical algorithms were chosen for comparison. RESULTS: The framework of a patient-based graph anomaly detection network for real-time quality control (PGADQC) was established and proven feasible for error detection. Compared with classic PBRTQC, the PGADQC showed a more balanced performance for both positive and negative biases. For different analytes, the average number of patient samples until error detection (ANPed) of PGADQC decreased variably, and reductions could reach up to approximately 95 % at a small bias of 0.02 taking calcium as an example. CONCLUSIONS: The PGADQC is an effective framework for patient-based quality control, integrating statistical and artificial intelligence algorithms. It improves error detection in a data-driven fashion and provides a new approach for PBRTQC from the data science perspective.

Machine learning-based nonlinear regression-adjusted real-time quality control modeling: a multi-center study.

OBJECTIVES: Patient-based real-time quality control (PBRTQC), a laboratory tool for monitoring the performance of the testing process, has gained increasing attention in recent years. It has been questioned for its generalizability among analytes, instruments, laboratories, and hospitals in real-world settings. Our purpose was to build a machine learning, nonlinear regression-adjusted, patient-based real-time quality control (mNL-PBRTQC) with wide application. METHODS: Using computer simulation, artificial biases were added to patient population data of 10 measurands. An mNL-PBRTQC was created using eight hospital laboratory databases as a training set and validated by three other hospitals' independent patient datasets. Three different Patient-based models were compared on these datasets, the IFCC PBRTQC model, linear regression-adjusted real-time quality control (L-RARTQC), and the mNL-PBRTQC model. RESULTS: Our study showed that in the three independent test data sets, mNL-PBRTQC outperformed the IFCC PBRTQC and L-RARTQC for all measurands and all biases. Using platelets as an example, it was found that for 20 % bias, both positive and negative, the uncertainty of error detection for mNL-PBRTQC was smallest at the median and maximum values. CONCLUSIONS: mNL-PBRTQC is a robust machine learning framework, allowing accurate error detection, especially for analytes that demonstrate instability and for detecting small biases.

Correlation Between the MTHFR C677T Genotype and Coronary Heart Disease in Populations from Gansu, China.

This study was designed to evaluate the relationship between polymorphisms in the methylenetetrahydrofolate reductase (MTHFR) gene and coronary heart disease (CHD) in populations from the Gansu region of China. The MTHFR C677T polymorphism genotypes from 209 patients with CHD, as confirmed by coronary angiography, and 212 non-CHD control patients were identified using PCR gold magnetic particle chromatography. We simultaneously evaluated homocysteine (Hcy) and folate levels in these samples using biochemical methods. The TT genotype of the MTHFR C677T locus was significantly more frequent in the CHD group than in the control, while the CC genotype was significantly less frequent in CHD patients than in non-CHD patients (p < 0.05). In addition, biochemical analysis revealed that the serum Hcy levels increased, and folate levels decreased in the TT genotype. Logistic regression analysis showed that this correlation was independent of nationality, sex, age, body mass index, medical history, and blood lipid level (p < 0.05). The occurrence of the TT genotype at the MTHFR C677T locus was closely associated with CHD in the Gansu population and may serve as a biomarker of increased risk for this disease.

A study on quality control using delta data with machine learning technique.

Background: In the big data era, patient-based real-time quality control (PBRTQC), as an emerging quality control (QC) method, is expanding within the clinical laboratory industry. However, the main issue of current PBRTQC methodology is data stability. Our study is aimed to explore a novel protocol for data stability by combining delta data with machine learning (ML) technique to improve the capacity of QC event detection. Methods: A data set of 423,290 laboratory results from Beijing Chao-yang Hospital 2019 patient results were used as a training set (n = 380960, 90%) and internal validation set (n = 42330, 10%). A further 22,460 results from Beijing Long-fu Hospital 2019 patient results were used as a test set. Three-type data (1) Single-type data processed by truncation limits; (2) delta-type data processed by truncation limits and (3)delta-type data processed by Isolated Forest (IF) algorithm were evaluated with accuracy, sensitivity, NPed, etc., and compared with previously published statistical methods. Results: The optimal model was based on Random Forest (RF) algorithm by using delta-type data processed by IF algorithm. The model had a better accuracy (0.99), sensitivity (0.99) specificity (0.99) and AUC (0.99) with the dependent test set, surpassing the critical bias of PBRTQC by over 50%. For the LYMPH#, HGB, and PLT, the cumulative MNPed of MLQC were reduced by 95.43%, 97.39%, and 97.97% respectively when compared to the best of the PBRTQC. Conclusion: Final results indicate that by integrating an innovative ML algorithm with the overall data processing protocol the detection of QC events is improved.

A multi-model fusion algorithm as a real-time quality control tool for small shift detection.

BACKGROUND: Patient-based real-time quality control (PBRTQC), a complement to traditional QC, may eliminate matrix effect from QC materials, realize real-time monitoring as well as cut costs. However, the accuracy of PBRTQC has not been satisfactory as physicians expect till now. Our aim is to set up a artificial intelligence-based QC for small error detection in real laboratory settings. Taking tPSA as our unique research subject, data extraction, data stimulation, data partition, model construction and evaluation were designed. METHODS: 84241 deidentified results for tPSA were extracted from Laboratory Information System of Aviation General Hospital. The data set was accumulated by way of data simulation. Independent training and test datasets were separated. After three classification models (RF, SVM and DNN) in ML constructed and weighted by information entropy, a multi-model fusion algorithm was generated. Performance of the fusion model was evaluated by comparing with optimal PBRTQC. RESULTS: For 4 PBRTQC methods, MovSO showed overall better performance for 0.2 µg/L bias and optimal MNPed was equal to 200. For the fusion model, MNPeds were less than 12 for all biases, and ACC surpassed MovSO nearly 100 times. Except for 0.01 µg/L bias, ACC was more than 0.9 for the rest of biases. FPR was apparently lower than MovSO, only 0.2% and 0.1%. CONCLUSION: The fusion model shows outstanding performance and reduces incorrect and omitting error detection, adaptable for the real settings.

Traceable machine learning real-time quality control based on patient data.

OBJECTIVES: Patient-based real-time quality control (PBRTQC) has gained attention as an alternative/integrative tool for internal quality control (iQC). However, it is still doubted for its performance and its application in real clinical settings. We aim to generate a newly and easy-to-access patient-based real-time QC by machine learning (ML) traceable to standard reference data with assigned values by National Institute of Metrology of China (NIM), and to compare it with PBRTQC for clinical validity evaluation. METHODS: For five representative biochemistry analytes, 1,195 000 patient testing results each were collected. After data processing, independent training and test sets were divided. Machine learning internal quality control (MLiQC) was set up by Random Forest in ML and was validated by way of both metrology algorithm traceability and 4 PBRTQC methods recommended by IFCC analytical working group. RESULTS: MLiQC were established. As an example of albumin (ALB) at the critical bias, the uncertainty of MLiQC was 0.14%, which was evaluated by standard reference data produced by NIM. Compared with four optimal PBRTQC methods at critical bias, the average of the number of patient samples from a bias introduced until detected (ANPed) of MLiQC averagely decreased from 600 to 20. The median and 95 quantiles of NPeds (MNPed and 95NPed) of MLiQC were superior to all optimal PBRTQCs above 90% for all test items. CONCLUSIONS: MLiQC is highly superior to PBRTQC and well-suited in real settings. The validation of the model from two aspects of algorithm traceability and clinical effectiveness confirms its satisfactory performance.

A highly accurate delta check method using deep learning for detection of sample mix-up in the clinical laboratory.

OBJECTIVES: Delta check (DC) is widely used for detecting sample mix-up. Owing to the inadequate error detection and high false-positive rate, the implementation of DC in real-world settings is labor-intensive and rarely capable of absolute detection of sample mix-ups. The aim of the study was to develop a highly accurate DC method based on designed deep learning to detect sample mix-up. METHODS: A total of 22 routine hematology test items were adopted for the study. The hematology test results, collected from two hospital laboratories, were independently divided into training, validation, and test sets. By selecting six mainstream algorithms, the Deep Belief Network (DBN) was able to learn error-free and artificially (intentionally) mixed sample results. The model's analytical performance was evaluated using training and test sets. The model's clinical validity was evaluated by comparing it with three well-recognized statistical methods. RESULTS: When the accuracy of our model in the training set reached 0.931 at the 22nd epoch, the corresponding accuracy in the validation set was equal to 0.922. The loss values for the training and validation sets showed a similar (change) trend over time. The accuracy in the test set was 0.931 and the area under the receiver operating characteristic curve was 0.977. DBN demonstrated better performance than the three comparator statistical methods. The accuracy of DBN and revised weighted delta check (RwCDI) was 0.931 and 0.909, respectively. DBN performed significantly better than RCV and EDC. Of all test items, the absolute difference of DC yielded higher accuracy than the relative difference for all methods. CONCLUSIONS: The findings indicate that input of a group of hematology test items provides more comprehensive information for the accurate detection of sample mix-up by machine learning (ML) when compared with a single test item input method. The DC method based on DBN demonstrated highly effective sample mix-up identification performance in real-world clinical settings.

Risk Factors of Difficult Pharyngeal Accidental Fishbones Ingestion.

OBJECTIVE: Accidental pharyngeal fishbone ingestion is a common complaint in ear, nose, and throat clinics. Approximately two-thirds of the accidentally ingested fishbones can be removed using tongue depressors and indirect laryngoscopy. However, the remaining third is challenging to identify and remove using these methods. These difficult fishbones require identification and removal via more advanced approaches. Video-guided laryngoscope is used to deal with difficult fishbones in our center. This study aimed to explore the risk factors for difficult fishbones. METHODS: A prospective study was performed at a teaching hospital on 2080 patients. Univariate and multivariate analyses were performed to identify the risk factors. RESULTS: The common fishbone locations were the tonsils (39.8%; defined as STEP-I), tongue base (37.1%), vallecula (13.3%; STEP-II), and hypopharynx (9.8%; STEP-III). With increasing STEP level, the ratio of difficult fishbones correspondingly increased (Z = 13.919, P < .001), and the proportions were 21.1%, 41.9%, and 70% in STEP-I, II, and III, respectively. In particular, fishbones in STEP-III (vs STEP-I) had a higher risk of difficult fishbones (odds ratio [OR]: 11.573, 95% CI: 7.987-16.769). Complaints of neck pain (yes vs no), foreign body sensation (yes vs no), and shorter length of fishbones always had a lower risk of difficult fishbones (OR: 0.455, 95% CI: 0.367-0.564; OR: 0.284, 95% CI: 0.191-0.422; OR: 0.727, 95% CI: 0.622-0.85). Missing teeth (yes vs no), swallowing behavior after fishbone ingestion (yes vs no), and male patients (vs female) had a higher risk of difficult fishbones (OR: 1.9, 95% CI: 1.47-2.456; OR: 1.631, 95% CI: 1.293-2.059; OR: 1.278, 95% CI: 1.047-1.56). CONCLUSIONS: Neck pain, foreign body sensation, fishbone length, patient age and sex, tooth status, and swallowing behavior after fishbone ingestion are independent risk factors for difficult fishbones.

Establishment of Multiple Myeloma Diagnostic Model Based on Logistic Regression in Clinical Laboratory.

BACKGROUND: Due to the insidious onset of multiple myeloma (MM), missed diagnosis and misdiagnosis have a serious impact on the health of MM patients. Simple, rapid, and valid laboratory screening is critical for MM clinical diagnosis. METHODS: We used routine laboratory tests to establish a simple, inexpensive, and non-invasive diagnostic model for MM based on logistic regression. In the retrospective analysis, a total of 273 newly diagnosed MM inpatients and 288 non-MM participants, from January 2016 to December 2018 in Beijing Chaoyang hospital, Capital Medical University, were divided into training set and validation set. Age, gender, and the related routine laboratory tests for MM, including albumin (ALB), globulin (GLB), lactate dehydrogenase (LDH), creatinine (Cr), calcium (Ca2+), hemoglobin (Hb) and platelet (PLT), were analyzed by multivariate logistic regression to develop a diagnostic model. RESULTS: A diagnostic model was calculated using the formula MM index=-((-18×gender-3×ALB-Hb)/10), based on the logistic regression. The MM index [22 (20 - 25)] of MM patients was significantly lower than that of non-MM [30 (29 - 31)] in the training set (p < 0.001). It showed an excellent diagnostic performance in diagnosing MM through a receiver operating characteristic (ROC) curve, and its corresponding sensitivity, specificity, and area under the curve (AUC) were 95.6%, 96.7%, and 0.982 (0.968, 0.997), respectively. At a diagnostic risk threshold of 28, the model identified MM with a sensitivity of 95.6% and a specificity of 98.1% by using independent validation data. There was a significant positive correlation (r = 0.845, p < 0.001) between the DS grading and the MM index among all the participants. CONCLUSIONS: The established diagnostic model of MM index can successfully identify newly diagnosed MM from healthy controls. The diagnostic model of MM index may also act as a predictor of the severity of MM without therapy.

Abnormal bone mineral density and bone turnover marker expression profiles in patients with primary spontaneous pneumothorax.

BACKGROUND: To examine the bone mineral density (BMD) and the role of bone biomarkers, including bone formation marker procollagen type I aminoterminal propeptide (PINP) and N-terminal midmolecule fragment osteocalcin (N-MID), bone resorption marker b-C-telopeptides of type I collagen (b-CTX) and tartrate-resistant acid phosphatase 5b (TRACP5b) in the pathogenesis of PSP. METHODS: Eighty-three consecutive primary spontaneous pneumothorax (PSP) patients (PSP group) and 87 healthy individuals (control group) were enrolled in this study. General data, including gender, age, height, weight, and body mass index (BMI), were recorded. Dual-energy X-ray absorptiometry, electrochemiluminescence immunoassay (ECLIA), and ELISA were used to evaluate bone mineral density and expression levels of bone metabolism markers, including PINP, b-CTX, TRACP5b, N-MID, and 25-hydroxyvitamin D (25-OH VD). RESULTS: Mean height was significantly greater in the PSP group compared with the control group, whereas weight and BMI were lower. Patients in the PSP group had significantly lower average bone mineral density, which mainly manifested as osteopenia (11/12, 91.7%); however, only one patient (8.3%) developed osteoporosis. Serum overexpression of PINP, b-CTX, TRACP5b, and N-MID were found in PSP patients. Expression of 25-OH VD was low in PSP patients. Bone resorption markers showed positive linear relationships with bone formation markers in all participants; whereas only TRACP5b expression negatively correlated with 25-OH VD. Expression levels of all bone turnover markers negatively correlated with BMI. Regression analysis identified risk factors of PSP as age, height, weight, and TRACP5b and 25-OH VD expression levels; whereas gender and PINP, b-CTX, and N-MID expression levels were not significantly associated with the onset of PSP. CONCLUSIONS: It had lower bone mineral density in PSP patients. Bone formation marker PINP, N-MID and bone resorption marker b-CTX, TRACP5b were upregulated in PSP patients. 25-OH VD expression was relatively low in this population of PSP patients. Age, height, weight, and expression levels of TRACP5b and 25-OH VD may be risk factors for PSP.

High prevalence of vitamin D deficiency in urban health checkup population.

BACKGROUND: Vitamin D deficiency is documented as a common health problem in the world. Limited data has been found on the prevalence of vitamin D deficiency in Beijing area. AIM: To investigate the prevalence s of vitamin D deficiency in urban Beijing residents and the seasonal and monthly serum 25(OH)D variation in this population. METHODS: This is an urban hospital based cross-sectional study lasting whole 2 years. 5531 (5-101 years old) urban Beijing residents for health checkup are recruited from December 9th, 2011 to December 8th, 2013. Each subject completed a questionnaire designed to quantify intake of vitamin D through food, vitamin D supplements, hours of sun exposure, sunscreen use over the past month. Serum 25(OH)D is statistically analyzed in accordance with gender, age, and time-lines. RESULTS: Vitamin D deficiency (Serum 25(OH)D level ≤20 ng/mL) and sever deficiency (Serum 25(OH)D level ≤ 10 ng/mL) are highly prevalent in this population. The prevalence of vitamin D deficiency is 87.1% and higher prevalence is found in female (89.0%) than male (84.9% P < 0.001). Severe vitamin D deficiency is also higher in female than male (59.3% and 42.7%, respectively, P < 0.001). Female under 20 and over 80 have lower 25(OH)D levels compared to 40-60 years old female (both P < 0.05). Severe vitamin D deficiency are also highly prevalence in this two group (60.9% and 54.1%) compared with 40-60 years old females (43.1%, both P < 0.05). Seasonal variation are also found in this population (P < 0.01). Autumn and summer have the higher 25(OH)D level than winter and spring in both genders (P < 0.001). Winter and spring have higher vitamin D deficiency and Severe deficiency than the other two seasons (P < 0.05). Serum 25(OH)D level peaks in October and troughed in April in both female and male. Lower serum 25(OH)D level are found in April than February (P < 0.05) in both gender. CONCLUSIONS: This is the first time to examine the prevalence of vitamin D deficiency among urban Beijing residents spanning the age spectrum. And Vitamin D deficiency and severe deficiency is found highly prevalent in this population, especially among females under 20 and older than 80 and in winter and spring seasons. Targeted prevention on vitamin D deficiency is urgent for this population.

Establishment and validation of serum free light chain reference intervals in an ethnic Chinese population.

BACKGROUND: Quantification of serum free light chains (FLC) and calculation of a kappa/lambda ratio using polyclonal antisera based immunoassays provide laboratories with a sensitive alternative to urine protein electrophoresis (UPE). However, the published 0.26 - 1.65 serum FLC kappa/lambda ratio reference intervals may not be suitable for different ethnic populations (such as the Han Chinese population presented) and require validation. This is particularly important where there are significant differences in ethnicity, and hence HLA prevalence, in the population studied. METHODS: Serum FLC reference intervals were determined using 326 Han Chinese blood donor volunteers. Sensitivities and specificities for the (i) serum FLC kappa/lambda ratio reference interval and (ii) UPE analyses were determined using 68 pre-treatment, serum immunofixation (sIFE) positive multiple myeloma (MM) patient samples, 54 sera from MM patients undergoing treatment, and 56 sIFE-negative samples from patients with no plasma cell dyscrasia (PCD). RESULTS: The 100% range for the serum FLC kappa/lambda ratio in this Han Chinese population was 0.32 - 1.52. Both Han Chinese blood donors and published kappa/lambda ratio reference ranges demonstrated higher diagnostic sensitivity and specificity for PCD than was seen with UPE. Highly abnormal serum FLC kappa/lambda ratios were observed in 68% of MM patients who had a negative UPE. Furthermore, a MM screening panel of SPE plus serum FLC assays achieved 100% diagnostic sensitivity compared to 97% for a UPE plus SPE algorithm. For MM patients undergoing therapy, 70% of UPE negative samples displayed an abnormal FLC ratio. CONCLUSIONS: This study confirms the requirement to verify normal FLC reference ranges in local populations. This Han Chinese reference range is narrower than the published range. FLC testing provides a robust, sensitive, and specific alternative to classic UPE assessment.

Spatiotemporal analysis of normal and pathological human vocal fold vibrations.

PURPOSE: For spatiotemporal analysis to become a relevant clinical tool, it must be applied to human vocal fold vibration. Receiver operating characteristic (ROC) analysis will help assess the ability of spatiotemporal parameters to detect pathological vibration. MATERIALS AND METHODS: Spatiotemporal parameters of correlation length and entropy were extracted from high-speed videos of 124 subjects, 67 without vocal fold pathology and 57 with either vocal fold polyps or nodules. Mann-Whitney rank sum tests were performed to compare normal vocal fold vibrations to pathological vibrations, and ROC analysis was used to assess the diagnostic value of spatiotemporal analysis. RESULTS: A statistically significant difference was found between the normal and pathological groups in both correlation length (P < .001) and entropy (P < .001). The ROC analysis showed an area under the curve of 0.85 for correlation length, 0.87 for entropy, and 0.92 when the 2 parameters were combined. A statistically significant difference was not found between the nodules and polyps groups in either correlation length (P = .227) or entropy (P = .943). The ROC analysis showed an area under the curve of 0.63 for correlation length and 0.51 for entropy. CONCLUSIONS: Although they could not effectively distinguish vibration of vocal folds with nodules from those with polyps, the spatiotemporal parameters correlation length and entropy exhibit the ability to differentiate normal and pathological vocal fold vibration and may represent a diagnostic tool for objectively detecting abnormal vibration in the future, especially in neurological voice disorders and vocal folds without a visible lesion.

Perturbation and nonlinear dynamic analysis of adult male smokers.

OBJECTIVE: Smoking results in a voice change, and the perception by smokers of an abnormal voice may encourage quitting behavior. Moreover, a disordered voice is often the first sign of vocal pathology. Efforts to evaluate voice have focused on classical acoustic analysis; however, nonlinear dynamic analysis has been shown to be a reliable objective method for the evaluation of voice. We compare the discriminatory ability of these two methods when applied to normal and smokers' voices. STUDY DESIGN: Prospective study. METHODS: The study included 73 subjects, 36 nonsmokers and 37 smokers. A segment of sustained vowel production was obtained from each subject. Acoustic dimension and correlation dimension (D2) analyses were applied to the data. Results were compared with a Mann-Whitney rank sum test, logistic regression, and receiver operating characteristics (ROC) analysis. RESULTS: D2 values for smokers were significantly higher than D2 values for nonsmokers (P<0.001). Jitter and shimmer analysis showed higher values for these parameters among smokers. Logistic regression indicated a higher predictive power with D2, and ROC analysis found no significant difference between the analysis methods. DISCUSSION: This study indicated that D2 is highly sensitive to changes associated with smoking and has the potential to be implemented clinically as an indicator of abnormal voice. Further research could focus on using nonlinear dynamic analysis to create a normative database, producing standards for monitoring voice changes caused by cigarette smoking.

Construction of a multiple myeloma diagnostic model by magnetic bead-based MALDI-TOF mass spectrometry of serum and pattern recognition software.

A diagnosis of multiple myeloma (MM) is difficult to make on the basis of any single laboratory test result. Accurate diagnosis of MM generally results from a number of costly and invasive laboratory tests and medical procedures. The aim of this work is to find a new, highly specific and sensitive method for MM diagnosis. Serum samples were tested in groups representing MM (n = 54) and non-MM (n = 108). These included a subgroup of 17 plasma cell dyscrasias, a subgroup of 17 reactive plasmacytosis, 5 B cell lymphomas, and 7 other tumors with osseus metastasis, as well as 62 healthy donors as controls. Bioinformatic calculations associated with MM were performed. The decision algorithm, with a panel of three biomarkers, correctly identified 24 of 24 (100%) MM samples and 46 of 49 (93.88%) non-MM samples in the training set. During the masked test for the discriminatory model, 26 of 30 MM patients (sensitivity, 86.67%) were precisely recognized, and all 34 normal donors were successfully classified; patients with reactive plasmacytosis were also correctly classified into the non-MM group, and 11 of the other patients were incorrectly classified as MM. The results suggested that proteomic fingerprint technology combining magnetic beads with MALDI-TOF-MS has the potential for identifying individuals with MM. The biomarker classification model was suitable for preliminary assessment of MM and could potentially serve as a useful tool for MM diagnosis and differentiation diagnosis.