Full Blood Count Data Research Articles

Accurate COVID-19 detection using full blood count data and machine learning

COVID-19 has spread rapidly worldwide in the past three years, triggering partial and full lockdowns globally. The successful control of the COVID-19 pandemic on a global scale depended heavily upon the accurate detection of COVID-19. However, the main diagnostic tests for COVID-19 have some significant limitations, e.g. the major nucleic acid (RT-PCR) tests while having a high sensitivity are time-consuming and require expensive equipment with the shortage of test kits in many countries. Antigen lateral flow tests have a lower sensitivity and they cannot be used during the early pandemic as well as usually more expensive than the full or complete blood count test used in this paper which can be potentially performed using a finger blood sample. The last decade has seen rapid growth of AI, particularly deep learning, which has found wide applications in medical image analysis, with results comparable to and even surpassing human expert performance. There have been several machine learning models reported for COVID-19 diagnostics or prognosis predictions, most of them based on CT and X-ray images. In this paper we have applied traditional machine learning and convolutional neural networks (CNNs) based deep learning to the blood test data obtained from hematology analyzers and demonstrated that the AI models can be used to detect COVID-19 with a high degree of accuracy (>97%). The performance of different classifiers will be compared and discussed. The work should have potential applications in current COVID-19 and future pandemics.

M019 Predicting erythrocyte sedimentation rate from full blood count data

M019 Predicting erythrocyte sedimentation rate from full blood count data

Assessing data quality from the Clinical Practice Research Datalink: a methodological approach applied to the full blood count blood test

A Full Blood Count (FBC) is a common blood test including 20 parameters, such as haemoglobin and platelets. FBCs from Electronic Health Record (EHR) databases provide a large sample of anonymised individual patient data and are increasingly used in research. We describe the quality of the FBC data in one EHR. The Test dataset from the Clinical Research Practice Datalink (CPRD) was accessed, which contains results of tests performed in primary care, such as FBC blood tests. Medical codes and entity codes, two coding systems used within CPRD to identify FBC records, were compared, with levels of mismatched coding, and number that could be rectified reported. The reliability of units of measurement are also described and missing data discussed. There were 14 entity codes and 138 medical codes for the FBC in the data. Medical and entity codes consistently corresponded to the same FBC parameter in 95.2% (n = 217,752,448) of parameters. In the 4.8% (n = 10,955,006) mismatches, the most common parameter rectified was mean platelet volume (n = 2,041,360) and 1,191,540 could not be rectified and were removed. Units of measurement were often either missing, partially entered, or did not appear to correspond to the blood value. The final dataset contained 16,537,017 FBC tests. Applying mathematical equations to derive some missing parameters in these FBCs resulted in 15 of 20 parameters available per FBC on average, with 0.3% of FBCs having all 20 parameters. Performing data quality checks can help to understand the extent of any issues in the dataset. We emphasise balancing large sample sizes with reliability of the data.

Components of the full blood count as risk factors for colorectal cancer detection: a systematic review protocol

IntroductionColorectal cancer is the fourth most common type of cancer and the second most common cause of cancer-related deaths in the UK. The full blood count (FBC) is a blood...

Do Selected Blood Inflammatory Markers Combined with Radiological Features Predict Proliferation Index in Glioma Patients?

The tumor microenvironment is partially characterized by a state of chronic inflammation, and radiologic features are related to the tumor's biological behavior. This study was conducted to explore whether peripheral blood inflammatory markers combined with radiologic features could predict proliferation potency. This study retrospectively reviewed 183 patients with a primary diagnosis of glioma. Clinical characteristics, preoperative peripheral full blood count data, and brain magnetic resonance imaging findings were reviewed to analyze the expression of inflammatory markers neutrophil lymphocyte ratio (NLR), monocyte lymphocyte ratio, and platelet lymphocyte ratio (PLR), as well as radiologic features such as location, peritumor edema, and contrast enhancement. Immunohistochemical staining was performed to determine the proliferation index (i.e., expression of Ki-67). Receiver operating characteristic curves for cutoff value, various bivariate tests, and binary logistic regression analyses were applied. Proliferation index was highly associated with tumor grade, showing a gradually increasing tendency. A Ki-67 cutoff value >9% predicted high-grade glioma (HGG). Mean NLR and PLR were significantly higher in the HGG group compared with the low-grade glioma group (NLR: 3.11 ± 0.59 vs. 4.27 ± 1.13; PLR: 133.07 ± 13.17 vs. 161.51 ± 38.99; P < 0.01 for both). Contrast enhancement was more likely in the HGG group, but there was no significant between-group difference in peritumor edema. Logistic regression analysis identified the following risk factors for prediction of proliferation potency: age, Karnofsky Performance Score, NLR, PLR, and contrast enhancement. However, age >43 years, NLR >3.68, and positive contrast enhancement independently predicted a higher proliferation rate. NLR and contrast enhancement were positively correlated with the proliferation potency of gliomas.

Evaluation of a prediction model for colorectal cancer: retrospective analysis of 2.5 million patient records.

Earlier detection of colorectal cancer greatly improves prognosis, largely through surgical excision of neoplastic polyps. These include benign adenomas which can transform over time to malignant adenocarcinomas. This progression may be associated with changes in full blood count indices. An existing risk algorithm derived in Israel stratifies individuals according to colorectal cancer risk using full blood count data, but has not been validated in the UK. We undertook a retrospective analysis using the Clinical Practice Research Datalink. Patients aged over 40 with full blood count data were risk‐stratified and followed up for a diagnosis of colorectal cancer over a range of time intervals. The primary outcome was the area under the receiver operating characteristic curve for the 18–24‐month interval. We also undertook a case–control analysis (matching for age, sex, and year of risk score), and a cohort study of patients undergoing full blood count testing during 2012, to estimate predictive values. We included 2,550,119 patients. The area under the curve for the 18–24‐month interval was 0.776 [95% confidence interval (CI): 0.771, 0.781]. Performance improves as the time interval reduces. The area under the curve for the age‐matched case–control analysis was 0.583 [0.574, 0.591]. For the population risk‐scored in 2012, the positive predictive value at 99.5% specificity was 8.8% with negative predictive value 99.6%. The algorithm offers an additional means of identifying risk of colorectal cancer, and could support other approaches to early detection, including screening and active case finding.

P167 Diagnostic accuracy of neutrophil-lymphocyte ratio in suspected paediatric inflammatory bowel disease

Background: Faecal calprotectin has an established role in the investigation of paediatric inflammatory bowel disease (PIBD) but may not be available. Availability is often hampered by non-provision of samples or delayed by long turnaround times. Exploration of other indices which may indicate a strong likelihood of PIBD in the absence of faecal biomarkers remains important. The neutrophil-lymphocyte ratio (NLR) can be calculated from a full blood count (FBC) and has been proposed as a marker of inflammation. We aimed to determine the diagnostic accuracy of NLR in suspected PIBD patients prior to their first endoscopic assessment. Methods: Children (aged <18 yrs) undergoing endoscopic assessment for suspected PIBD were identified from our prospective regional IBD database and local endoscopy lists. Review of laboratory records identified the FBC performed closest and prior to endoscopy. Patients were eligible for inclusion if complete FBC data was available within 6 months prior to endoscopy, during the initial diagnostic cycle. Statistical analyses were performed using Vassarstats and EasyROC v1.3. Results: 182 patients met the inclusion criteria; 89 cases of PIBD and 93 cases with non-IBD diagnoses or normal investigations. The PIBD group were older (12.6yr vs 8.8yr; p<0.0001) with no difference in sex ratio (p=0.575). Median time from blood sample to endoscopy was 16 days (IQR 0–49) and was lower in the PIBD group (4d vs 37d; p<0.0001). NLR was significantly higher in PIBD (median 3.04, IQR 2.03–4.18) than in non-IBD patients (1.4, 0.95–2.03) (p<0.0001). Area under the receiver operator curve (AUROC) for NLR as a diagnostic test for PIBD was 0.81 (95% CI 0.75–0.88). Using an optimal cut-off of 2.37 gave a sensitivity of 67% (95% CI 57–77) and specificity 85% (95% CI 76–92). In patients with otherwise normal blood markers (CRP, ESR, albumin, WCC, platelets), median NLR remained significantly higher in PIBD (2.32 [1.74–3.28] vs 1.43 [1.01–1.97]; p<0.001, AUROC 0.77 [95% CI 0.65–0.89]). Using any abnormal test from CRP, ESR, albumin and NLR, sensitivity was 86% (95% CI 77–92), specificity 66% (95% CI 54–76), positive predictive value 74% (95% CI 64–83) and negative predictive value 80% (95% CI 68–89). Conclusions: NLR is significantly higher in patients with PIBD compared to cases with non-IBD diagnoses or normal investigations. This remains the case in patients with normal inflammatory markers. NLR provides moderate sensitivity and specificity for the diagnosis of PIBD and sensitivity is improved by combination with other readily available blood parameters. We propose that NLR has a useful place in the diagnostic work-up of patients with PIBD, especially in the resource-poor setting and where there are barriers to obtaining faecal samples.

PWE-094 The Prevalence of Anaemia in Inflammatory Bowel Disease in Relation to Disease Activity, as Stratified by Faecal Calprotectin

IntroductionThe prevalence of anaemia in IBD varies significantly between published studies, ranging from 6–74%. However, underlying disease activity, a potential explanation for this variability, has not been accurately correlated with...

Mo1260 The Prevalence of Anaemia in Inflammatory Bowel Disease in Relation to Disease Activity, As Stratified by Faecal Calprotectin

Introduction The prevalence of anaemia in IBD varies significantly between published studies, ranging from 6–74%. However, underlying disease activity, a potential explanation for this variability, has not been accurately correlated with the frequency of anaemia to date. Faecal calprotectin (FC) is a surrogate marker of underlying mucosal inflammation. The objective of this study was to investigate the prevalence of anaemia and its correlation with disease activity in IBD, in a large cohort of patients with matched full blood count (FBC) and FC data. Methods Patients with confirmed IBD from the Edinburgh faecal calprotectin registry (EFCR) were identified. Where multiple FCs were available, the most recent result was taken as reference. Blood test results were obtained from the electronic record covering a period one month either side of the FC. The WHO criteria was used to define anaemia (Hb Results 1226 patients (771 CD, F = 65%, 455 UC, F = 35%) with matched FC and FBC data were analysed. The median age was 44y (IQR 31–57), median disease duration 102 months (IQR 31–207). Overall, 314/1226 patients (25.6%) were anaemic, 185/314 (58.9%) of which were female. Anaemia was observed more frequently in patients with active as opposed to inactive CD (110/328 [33.5%] vs 65/443 [14.7%], p Conclusion In this cohort over 25% of patients with IBD were anaemic. There is a clear correlation between disease activity and anaemia in both CD and UC, but this is unrelated to disease distribution in CD. Anaemia in asymptomatic patients should alert clinicians to the possibility of subclinical active mucosal inflammation. These data and the ROC analysis provide further support for optimising disease treatment in IBD, targeting a FC level of Disclosure of Interest None Declared.

Donor lymphocyte collections using the spectra Optia MNC version 5

Methods and Materials: 10 non-mobilized allogeneic donors have had DLI apheresis collections on Optia MNC version 5. Collection preference was set to deliver a low haematocrit final product. ACD-A anticoagulation was used and apheresis completed at an AC ratio of 13.5 and infusion rate of 0.8mL/min/L total blood volume (TBV). Run data were recorded collected from the Spectra Optia apheresis reporting system at the end of each procedure. Full blood count data and CD3+/CD19+ target cell numbers were completed for the donor before and after apheresis and from the final product. Performance was measured by target cell yield and collection efficiency together with final product volume and haematocrit. Platelet loss to the donor was also measured after apheresis.

Post-transfusion white cell count in the sick preterm neonate.

A previous report demonstrated post-transfusion leucocytosis as a potential confounding factor in the diagnosis of sepsis in critically ill adult patients. In We wished to establish if the same phenomenon occurred in the sick preterm neonate and whether this significantly altered the indices considered for potential neonatal infection. Transfusion and full blood count data in a level 3 neonatal intensive care unit were prospectively recorded for 3 months. One hundred and fourteen transfusion events were recorded from 37 infants. Median white blood cell count increased 0.9 x 109/L (confidence interval (CI) 0.4-2.4) in the first 8 h following transfusion (P = 0.032). Median neutrophil count increased by 0.4 x 109/L (CI 0.1-1.7) in the same 8 h (P = 0.05). Median neutrophil left shift decreased 1.2% (CI 1.1-5.8%) over the 24 h post-transfusion. No change in band count was observed. A mild post-transfusion white cell increase occurs in sick neonates. Because of the magnitude of effect, it is unlikely that this interesting physiological response would interfere with the diagnosis of sepsis in this population.

Chronic lymphatic leukaemia: An investigation of HLA antigen frequencies and white cell differential counts in patients, relatives and controls

Histocompatibility antigen (HLA) frequencies in chronic lymphatic leukaemia (CLL) patients and control subjects were compared with respect to disease susceptibility and prognosis. Additionally, HLA and full blood count data were compared in relatives of 25 patients and 31 controls. We found no association of HLA with susceptibility although the presence of HLA B12, alone or in combination with HLA A2, indicated better prognosis. Relatives HLA identical with the patients showed no evidence of white cell disorder when compared with haplo- or non-identical relatives, or controls. As a group, however, relatives of patients had fewer lymphocytes than relatives of controls.