Noninvasive Hemoglobin Measurement Research Articles

Non-invasive hemoglobin measurement using embedded platform

This paper is based on the design and development of an in-house algorithm system to Estimate Total Hemoglobin in human blood using the Arduino based Embedded system. In our present society, monitoring Hemoglobin of subjects is very important in hospitals to keep track of low/high hemoglobin count during operations, delivery, dialysis, blood donations, perform stress test, monitoring patients in intensive care units, etc. is often done using invasive methods. In our approach, we have designed a finger probe using five LEDs on a multi-chip having wavelengths 670 nm/770 nm/810 nm/850 nm and 950 nm, and a single silicon photo detector with an inbuilt trans-impedance amplifier to acquire the signals through the finger using the Photoplethysmography (PPG) principle. The non-invasive method can be replaced in hospitals and homes to monitor the total hemoglobin level which is in real-time, painless, and infection-free from needles. We have tested our system on 8 subjects for three and five wavelengths and the values were compared to the hemoglobin measured using invasive methods in the pathology laboratory. The result demonstrates that the five wavelengths PPG method significantly improves the measurement accuracy of total hemoglobin, reducing the mean absolute error between the reference and the estimated hemoglobin values from 1.3778 g/dL for three wavelengths PPG to 0.3499 g/dL for five wavelengths PPG. The five wavelengths PPG system designed showed good linearity to estimate hemoglobin with the Regression Coefficient (R2 = 0.964) compared to three wavelengths PPG system with the Regression Coefficient (R2 = 0.688).

The usefulness of non-invasive co-oximetry haemoglobin measurement for screening pre-operative anaemia.

Pre-operative anaemia (haemoglobin<13.0g.dl-1 ) is a modifiable peri-operative risk-factor. This is screened for using formal laboratory testing. A non-invasive finger-probe sensor that can accurately measure haemoglobin is a possible alternative. This study considers the accuracy of non-invasive haemoglobin measurement using the Rad-67™ Rainbow (Masimo Corp., Irvine, CA, USA) compared with formal laboratory testing and its usefulness in detecting pre-operative anaemia. A total of 392 patients had measurements taken for non-invasive haemoglobin and perfusion index values using the Rad-67 Rainbow, alongside further peri-operative parameters and a formal laboratory haemoglobin test. Bland-Altman and sensitivity analysis showed that the limits of agreement between non-invasive and formal laboratory haemoglobin testing were between -1.95g.dl-1 and 2.23g.dl-1 (p<0.001). The overall performance of non-invasive haemoglobin measurement was better in men than women (ROC 91.1% vs. 78.2%) and less biased in men, mean -0.08 (SD 1.09, 95%Cl -0.23-0.07) compared with women (mean 0.38 (SD 0.99, 95%CI 0.24-0.52)). Pre-operative anaemia was more prevalent in women than men (50.3% vs. 14.4%). The sensitivity of non-invasive anaemia detection (haemoglobin<13g.dl-1 ) was 66% for women and 52% for men. A non-invasive haemoglobin value of 14.0g.dl-1 had an overall 91% sensitivity for detecting pre-operative anaemia (82% in men and 93% in women). The Rad-67 Rainbow is inadequate for the estimation of formal laboratory haemoglobin and lacks sensitivity for detecting pre-operative anaemia, especially in women. Further advancement in technology and accuracy is needed before it can be recommended as a routine pre-operative screening test.

Transcutaneous Hemoglobin Screening in an Adult Orthopaedic Trauma Population.

To evaluate a noninvasive hemoglobin measurement device in an orthopaedic trauma population. Prospective. Level 1 trauma center. One hundred five patients consecutively admitted to the orthopaedic trauma service after surgical treatment of fracture. Transcutaneous hemoglobin (TcHgb) monitoring using the Masimo Pronto Pulse CO-Oximeter model with Rainbow SET Technology for spot TcHgb measurement. TcHgb measurements and standard venipuncture hemoglobin (vHgb) were obtained. Patient preferences for each were recorded. TcHgb measurements were obtained in 100 patients and compared with their corresponding vHgb measurements. The mean vHgb and TcHgb were 10.2 ± 1.9 g/dL and 11.2 ± 2.1 g/dL, respectively, and the mean difference was 1.1 ± 1.6 g/dL, which was statistically different from 0 (P < 0.001). In 76% of cases, the TcHgb device overestimated vHgb. In a subgroup of patients undergoing procedures with minimal expected blood loss (external fixators of knee or ankle, irrigation and debridement, or open reduction and internal fixation of ankle or calcaneal fractures), the mean difference between vHgb and TcHgb was 0.68 ± 1.6 g/dL (P = 0.06). A preliminary study of TcHgb monitoring with the tested device as a potential screening mechanism to limit unnecessary blood draws showed statistical difference from vHgb; however, the mean bias 1.1 g/dL of hemoglobin was notably small. In a subgroup of patients undergoing procedures with minimal expected blood loss, the device may have merit. Larger studies are required to determine the clinical relevance of differences in measurements between the 2 methods. Diagnostic Level II. See Instructions for Authors for a complete description of levels of evidence.

Clinical implications of using non-invasive haemoglobin monitoring for red blood cell transfusion decision in hip arthroplasty

IntroductionThe decision to transfuse red blood cells requires accurate haemoglobin concentration values. In this study, we evaluated if continuous non-invasive haemoglobin (SpHb) measurement could substitute laboratory determined haemoglobin (LabHb) in patients undergoing elective hip replacement. As secondary objective, we analyzed the trend of the difference between techniques. Materials/methodsLabHb measurements were done using an automated analyser and SpHb measurements were acquired using Radical-7®. In randomly selected patients undergoing hip replacement, whenever blood was collected for LabHb, concomitant SpHb was recorded. Correlation, bias and accuracy of SpHb were calculated in comparison with LabHb. Results108 paired measurements were obtained from 43 patients. The Pearson R of the correlation between SpHb and LabHb was 0.7 (p < 0.001). Bland-Altman test revealed a bias of 1 ± 1.4 g dL−1, meaning Lab Hb was recurrently higher than SpHb. Limits of agreement were [−1.7; 3.8]. Considering RBC transfusion threshold of 8 g dL−1, we found that in two situations transfusion decision would differ based on the measurement considered. Trending ability of SpHb study showed a significant difference between preoperative and postoperative LabHb-SpHb. DiscussionThere was a good correlation between SpHb and LabHb, while bias and limits of agreement were higher than those in literature. There was a limited trending ability of SpHb during the perioperative period. Despite this, using SpHb instead of LabHb for decision making regarding transfusion would only change the decision in 1.9 % of our cases. Our findings suggest that this device could be used as a reference but cannot replace venous puncture as gold standard.

Evaluation of noninvasive hemoglobin monitoring in children with congenital heart diseases.

Noninvasive measurement of blood hemoglobin could save time and decrease the risk of anemia and infection. The accuracy of CO-oximetry-derived noninvasive hemoglobin (Sp-Hb) had been evaluated in pediatric population; however, its accuracy in children with congenital heart disease has not been studied till date. We evaluated the accuracy of Sp-Hb in relation to laboratory-measured hemoglobin (Lab-Hb) in children with congenital heart disease. This prospective observational study included children with congenital heart disease undergoing procedural intervention. Sp-Hb measurements were obtained using Radical-7 Masimo pulse CO-oximeter and were compared against simultaneous Lab-Hb measurements obtained from the arterial line. Children were divided in cyanotic and acyanotic, and separate analysis was performed for each group. The values of both measurements were analyzed using Spearman's correlation coefficient and Bland-Altman analysis. Correlation was performed between Sp-Hb and Lab-Hb bias and each of arterial oxygen saturation and perfusion index. One-hundred and eleven pairs of readings were obtained from 65 children. The median (quartiles) age and weight of the children were 1 (1.2-4)years and 11 (8-17)kg, respectively. There was moderate correlation between Lab-Hb and Sp-Hb with a correlation coefficient (95% confidence interval [CI]) of 0.75 (0.63-0.83) in acyanotic children and 0.62 (0.37-0.79) in cyanotic children. The mean bias (95% limits of agreements) was -0.4g/dL (-2.4 to1.6g/dL) and 1g/dL (-2.7to4.6g/dL) in acyanotic and cyanotic children, respectively. The mean bias between Sp-Hb and Lab-Hb showed a weak negative correlation with oxygen saturation (r [95% CI]): (-0.36 [-0.51--0.18]), and a weak positive correlation with the perfusion index (r [95% CI]): (0.19 [0.01-0.37]). The large bias and the wide limits of agreement between Sp-Hb and Lab-Hb denote that Masimo-derived Sp-Hb is not accurate in children with congenital heart disease especially in the cyanotic group; the error in Sp-Hb increases when oxygen saturation decreases.

Non-invasive haemoglobin measurement among pregnant women using photoplethysmography and machine learning

Estimation of haemoglobin(Hb) using non-invasive methods has caught wide interest among researchers all over the globe. Among the various devices and methods adopted for non-invasive haemoglobin (SpHb) measurement, prediction of SpHb using PPG can provide earlier and faster diagnosis. The primary objective and motivation of this study is to assess the efficiency and effectiveness of SpHb monitoring among pregnant women using PPG and generalized linear regression technique. PPG signal was acquired from pregnant women with prior consent and Hb was predicted from the time-domain attributes of PPG. Hb value as calculated by the invasive lab methods was compared with the predicted SpHb value. The absolute bias between the SpHb predicted and Hbref was 0.73 g/dL (SD 0.62). To analyse and evaluate the performance of the proposed word, correlation coefficient between SpHb and Hblab, was calculated using IBM SPSS software.

Efficacy of noninvasive hemoglobin measurement by pulse co-oximetry in neonates

Background: Hemoglobin (Hb) measurement is one of the most commonly performed laboratory tests in neonates. The advent of new technologies enables us to carry out noninvasive Hb measurement, without having to prick the neonate, whereby reducing iatrogenic blood loss. There have been only few studies trying to ascertain the relationship between noninvasive Hb and conventional Hb in neonates. Against this background, this study was undertaken to evaluate the efficacy of noninvasive spectrophotometric Hb (SpHb) measurement by the pulse co-oximetry in neonates in comparison to laboratory Hb (Lab-Hb) and to study the influence of factors such as level of Hb and serum bilirubin on same. Materials and Methods: A cross-sectional study was carried out in the postnatal ward and neonatal intensive care unit of a tertiary hospital in Southern India. Hemodynamically stable admitted neonates (n = 100) had their SpHb estimation done using the Masimo Radical-7 Pulse Co-oximeter at the time they were being sampled for Hb. The paired data was then analyzed for the relationship between SpHb and Lab-Hb. Results: The mean ± standard deviation g/dL for Lab-Hb and SpHb was 16.21 ± 1.92 and 15.45 ± 1.28, respectively. There was a reasonable positive correlation between Lab-Hb and SpHb. (Pearson correlation: R = 0.714; P

Non-invasive Estimation of Haemoglobin Level Using PCA and Artificial Neural Networks

Objective: Haemoglobin(Hb) measurement is generally performed by the traditional “fingerstick” test i.e., by invasively drawing blood from the body. Although the conventional laboratory measurement is accurate, it has its own limitations such as time delay, inconvenience of the patient, exposure to biohazards and the lack of real-time monitoring in critical situations. Non-invasive Haemoglobin Measurement (SpHb) has gained enormous attention among researches and can provide an earlier diagnosis to polycythemia, anaemia, various cardiovascular diseases, etc. Currently, Photoplethysmograph signal (PPG) is used for measuring oxygen saturation, to monitor the depth of anesthesia, heart rate and respiration monitoring. But through detailed statistical analysis, PPG signal can provide further information about various blood components. Investigation / Methodology: In this paper, an approach for non-invasive measurement of Hb using PPG, Principal Component Analysis (PCA) and Neural Network is proposed. A transmissive type PPG sensor is developed which is interfaced with Crowduino for the acquisition of PPG. From the obtained PPG signal, Principal Components (PC) are extracted. SpHb is predicted followed by the extraction of features from the PC. The analysis involves the SpHb prediction using a single PC, double PC and finally all the three PC. The predicted SpHb is evaluated with Hblab in terms of R-value, Mean Absolute Error, Mean Squared Error and Root Mean Squared Error. Conclusion: An approach for non-invasive measurement of Hb using Principal Components obtained from the PPG signal is discussed. The SpHb value is compared with the Hblab values. Correlation R-value between SpHb and Hblab is 0.77 when three principal components are used. Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE) between SpHb and Hblab are 0.3, 0.44 and 0.6633 respectively when SpHb is measured with three principal components. It is evident from the result analysis that SpHb shows the promising result when all the three principal components are used. However, one of the limitations of the work is that the population setting chosen for the work does not include paediatric patients, accurately ill patient, pregnant population and surgical patients. With detailed analysis on a wide range of population setting, Hb prediction using PPG is a promising approach for non-invasive measurement.

Non-Invasive Estimation of Hemoglobin Using a Multi-Model Stacking Regressor.

We describe a novel machine-learning based method to estimate total Hemoglobin (Hb) using photoplethysmograms (PPGs) acquired non-invasively. In a study conducted in Karnataka, India, 1583 women (pregnant and non-pregnant) of childbearing age, with Hb values ranging between 1.6 to 14.8g/dL, had their Hb values estimated using intravenous blood samples and concurrently by a finger sensor custom designed and prototyped for this study. The finger sensor collected PPG signals at four wavelengths: 590nm, 660nm, 810nm, and 940nm. A novel feature vector was derived from these PPGs. A machine learning model comprising of a two-layer stack of regressors including Least Absolute Shrinkage and Selection Operator (LASSO), Ridge, Elastic Net, Adaptive (Ada) Boost and Support Vector Regressors (SVR) was designed and tested. We report a statistically significant Pearson's correlation coefficient (PCC) of 0.81 (p < 0.01) between the Hb value estimated by the proposed methodology and gold standard values of Hb, with a Root Mean Square Error (RMSE) of 1.353 ± 0.042g/dL. The performance of the stacked regressor model was significantly better than the performance of individual regressors (low RMSE, and better CC; p < 0.05). Post-hoc analysis showed that including pregnant women in the training data set significantly improved the performance of the algorithm. This article demonstrates the feasibility of a machine learning based non-invasive hemoglobin measurement system, especially for maternal anemia detection. By developing and demonstrating a machine learning approach on a large data set, we have demonstrated that such an approach could become the basis for a public health screening tool to detect and treat maternal anemia and could supplement global health intervention strategies.

Validity of accuracy and trending ability of non-invasive continuous total hemoglobin measurement in complex spine surgery: a prospective cohort study

BackgroundPatients undergoing complex spine surgery present with multilevel spinal involvement, advanced age, and multiple comorbidities. Surgery is associated with significant blood loss and remarkable hemodynamic changes. The present study aimed to investigate the accuracy and trending ability of a non-invasive continuous method to monitor hemoglobin (SpHb) concentrations using a Radical-7™ Pulse CO-Oximeter in complex spine surgery.MethodsForty-nine patients who underwent complex spine surgery were enrolled in this prospective observational study. Multiple time points were established for data collection throughout the operation. Simultaneous SpHb–total hemoglobin (tHb) paired data were recorded for analyses. Linear regression analysis, Bland–Altman plot, four-quadrant plot, and Critchley polar plot were used to assess the accuracy and trending ability of the monitor.ResultsA total of 272 pairs of SpHb-tHb data were available and were divided into two groups based on the perfusion index (PI): PI values ≥1.0 (n = 200) and PI values < 1.0 (n = 72). The correction coefficients (r) between SpHb and tHb were 0.6946 and 0.6861 in the groups with PI values ≥1.0 and < 1.0, respectively (P < 0000.1). In the ≥1.0 group, the mean bias was − 0.21 g/dL and the percentage error (PE) was 15.85%, whereas in the < 1.0 group, the mean bias was − 0.04 g/dL and the PE was 17.42%. Four-quadrant plot revealed a concordance rate of 85.11%, whereas the Critchley polar plot showed a concordance rate of 67.21%.ConclusionsThe present study demonstrates the acceptable accuracy of the Radical-7™ Pulse CO-Oximeter even with a low PI. However, the trending ability was limited and unsatisfactory.

Assessment of haemoglobin measurement by several methods - blood gas analyser, capillary and venous HemoCue® , non-invasive spectrophotometry and laboratory assay - in term and preterm infants.

A laboratory haematology analyser is the gold standard for measuring haemoglobin concentration but has disadvantages, especially in neonates. This study compared alternative blood-sparing and non-invasive methods of haemoglobin concentration measurement with the gold standard. Haemoglobin concentrations were measured using a laboratory haematology analyser (reference method), blood gas analyser, HemoCue® using venous and capillary blood samples and a newly developed non-invasive sensor for neonates < 3 kg. A total of 63 measurements were performed. Body weight (2190 (1820-2520 [967-4450]) g) and haemoglobin concentration (12.3 (10.6-15.2 [8.2-20.5]) g.dl-1 ) varied widely. Bias/limits of agreement between the alternative methods and reference method were -0.1/-1.2 to 1.0 g.dl-1 (blood gas analyser), -0.4/-1.8 to 1.1 g.dl-1 (HemoCue, venous blood), 0.7/-1.9 to 3.2 g.dl-1 (HemoCue, capillary blood) and -1.2/-4.3 to 2 g.dl-1 (non-invasive haemoglobin measurement). Perfusion index, body weight and fetal haemoglobin concentration did not affect the accuracy of the alternative measurement methods, and these were successfully applied in term and preterm infants. However, the accuracies of non-invasive haemoglobin measurement and HemoCue of capillary blood especially lacked sufficient agreement with that of the reference method to recommend these methods for clinical decision making.

The Clinical Utility of Noninvasive Pulse Co-oximetry Hemoglobin Measurements in Dark-Skinned Critically Ill Patients.

The primary objective of this study was to assess the clinical usefulness of a point-of-care device which measures hemoglobin noninvasively (SpHb) in a group of critically ill participants with dark skin pigmentation. One hundred forty-six adult and pediatric participants from a multidisciplinary intensive care unit had intermittent readings of noninvasive hemoglobin measurements performed at a minimum of 4 hourly intervals. A total of 371 readings were analyzed. Concurrent blood samples were taken to assess hemoglobin levels using point-of-care blood gas analyzer, as well as sent to a central laboratory where hemoglobin was measured using the sodium lauryl sulfate method. Bland-Altman plots were constructed to assess the agreement between results from the 2 point-of-care devices with the reference standard (laboratory hemoglobin). SpHb exhibited significant bias when compared to laboratory hemoglobin, while blood gas hemoglobin did not. Mean bias for SpHb was +1.64 with limits of agreement of -1.03 to 4.31 compared to blood gas hemoglobin which showed a bias of 0.26 and limits of agreement of -0.84 to 1.37. The magnitude of the bias for SpHb increased with increasing mean hemoglobin levels. Of all the additional study variables assessed for effect on the bias, only Acute Physiology and Chronic Health Evaluation II score in adult patients (P < .0001) and mean arterial blood pressure (P = .001) had an effect. Skin pigmentation did not have any effect on the magnitude of bias. Noninvasive Hemoglobin measurement is a promising tool in dark-skinned critically ill patients with low hemoglobin levels, but requires further refinements for it to have clinical usefulness.

Economic Analysis of the Reduction of Blood Transfusions during Surgical Procedures While Continuous Hemoglobin Monitoring Is Used.

Background: Two million transfusions are performed in Spain every year. These come at a high economic price for the health system, increasing the morbidity and mortality rates. The way of obtaining the hemoglobin concentration value is via invasive and intermittent methods, the results of which take time to obtain. The drawbacks of this method mean that some transfusions are unnecessary. New continuous noninvasive hemoglobin measurement technology can save unnecessary transfusions. Methods: A prospective study was carried out with a historical control of two homogeneous groups. The control group used the traditional hemoglobin measurement methodology. The experimental group used the new continuous hemoglobin measurement technology. The difference was analyzed by comparing the transfused units of the groups. The economic savings was calculated by multiplying the cost of a transfusion by the difference in units, taking into account measurement costs. Results: The percentage of patients needing a transfusion decreased by 7.4%, and the number of transfused units per patient by 12.56%. Economic savings per patient were €20.59. At the national level, savings were estimated to be 13,500 transfusions (€1.736 million). Conclusions: Constant monitoring of the hemoglobin level significantly reduces the need for blood transfusions. By using this new measurement technology, health care facilities can significantly reduce costs and improve care quality.

Accuracy and trending of non-invasive hemoglobin measurement during different volume and perfusion statuses.

The evolution of non-invasive hemoglobin measuring technology would save time and improve transfusion practice. The validity of pulse co-oximetry hemoglobin (SpHb) measurement in the perioperative setting was previously evaluated; however, the accuracy of SpHb in different volume statuses as well as in different perfusion states was not well investigated. The aim of this work is to evaluate the accuracy and trending of SpHb in comparison to laboratory hemoglobin (Lab-Hb) during acute bleeding and after resuscitation. Seventy patients scheduled for major orthopedic procedures with anticipated major blood loss were included. Radical-7 device was used for continuous assessment of SpHb, volume status [via pleth variability index (PVI)] and perfusion status [via perfusion index (PI)]. Lab-Hb and SpHb were measured at three time-points, a baseline reading, after major bleeding, and after resuscitation. Samples were divided into fluid-responsive and fluid non-responsive samples, and were also divided into high-PI and low-PI samples. Accuracy of SpHb was determined using Bland-Altman analysis. Trending of SpHb was evaluated using polar plot analysis. We obtained 210 time-matched readings. Fluid non-responsive samples were 106 (50.5%) whereas fluid responsive samples were 104 (49.5%). Excellent correlation was reported between Lab-Hb and SpHb (r = 0.938). Excellent accuracy with moderate levels of agreement was also reported between both measures among all samples, fluid non-responsive samples, fluid-responsive samples, high-PI samples, and low-PI samples [Mean bias (limits of agreement): 0.01 (- 1.33 and 1.34) g/dL, - 0.08 (- 1.27 and 1.11) g/dL, 0.09 (- 1.36 and 1.54) g/dL, 0.01 (- 1.34 to 1.31) g/dL, and 0.04 (- 1.31 to 1.39) g/dL respectively]. Polar plot analysis showed good trending ability for SpHb as a follow up monitor. In conclusion, SpHb showed excellent correlation with Lab-Hb in fluid responders, fluid non-responders, low-PI, and high PI states. Despite a favorable mean bias of 0.01g/dL for SpHb, the relatively wide levels of agreement (- 1.3 to 1.3g/dL) might limit its accuracy. SpHb showed good performance as a trend monitor.

Lessons Learned From Field Deployment of the Hemaapp: A Non-Invasive Hemoglobin Measurement tool

Lessons Learned From Field Deployment of the Hemaapp: A Non-Invasive Hemoglobin Measurement tool

Non-invasive Hemoglobin Measurement for Anemia Diagnosis

Non-invasive Hemoglobin Measurement for Anemia Diagnosis

Can noninvasive hemoglobin measurement reduce the need for preoperative venipuncture in pediatric outpatient surgery?

Noninvasive measurements of hemoglobin in the pediatric perioperative setting could be helpful to avoid venipunctures in children. The present study aims to evaluate this by using a noninvasive device for hemoglobin determination. We compared noninvasively obtained hemoglobin with laboratory hemoglobin concentrations in children during their preoperative assessment. In an observational study, 122 nonanemic children (age 4.2±1.6years) who were scheduled to undergo different surgical procedures under general anesthesia were included. In their preoperative preparations, single invasive blood samples for laboratory hemoglobin concentrations were routinely taken following hospital policy and compared to simultaneous noninvasive determinations of hemoglobin. A preoperative invasive value ≤9g/dL would have caused cancelation of surgery and implied further investigations. A Bland-Altman plot showed that the average difference between noninvasively obtained hemoglobin and laboratory hemoglobin concentration was -0.44g/dL (bias) with a standard deviation of the mean bias of 1.04g/dL. A hemoglobin error grid showed that the noninvasive device could identify almost all invasive hemoglobin values >9g/dL. In total, there were 4 false-positive values where noninvasively obtained hemoglobin observations were below while the paired invasive values were above 9g/dL. The data in this pediatric setting suggest that the device may eliminate the need for venipuncture in nonanemic children.

Optical wavelength selection for portable hemoglobin determination by near-infrared spectroscopy method

Hemoglobin concentration is commonly used in clinical medicine to diagnose anemia, identify bleeding, and manage red blood cell transfusions. The golden standard method for determining hemoglobin concentration in blood requires reagent. Spectral methods were advantageous at fast and non-reagent measurement. However, model calibration with full spectrum is time-consuming. Moreover, it is necessary to use a few variables considering size and cost of instrumentation, especially for a portable biomedical instrument. This study presents different wavelength selection methods for optical wavelengths for total hemoglobin concentration determination in whole blood. The results showed that modelling using only two wavelengths combination (1143nm, 1298nm) can keep on the fine predictability with full spectrum. It appears that the proper selection of optical wavelengths can be more effective than using the whole spectra for determination hemoglobin in whole blood. We also discussed the influence of water absorptivity on the wavelength selection. This research provides valuable references for designing portable NIR instruments determining hemoglobin concentration, and may provide some experience for noninvasive hemoglobin measurement by NIR methods.

HemaApp

We present HemaApp, a smartphone application that noninvasively monitors blood hemoglobin concentration using the smartphone's camera and various lighting sources. Hemoglobin measurement is a standard clinical tool commonly used for screening anemia and assessing a patient's response to iron supplement treatments. Given a light source shining through a patient's finger, we perform a chromatic analysis, analyzing the color of their blood to estimate hemoglobin level. We evaluate HemaApp on 31 patients ranging from 6 -- 77 years of age, yielding a 0.82 rank order correlation with the gold standard blood test. In screening for anemia, HemaApp achieve a sensitivity and precision of 85.7% and 76.5%. Both the regression and classification performance compares favorably with our control, an FDA-approved noninvasive hemoglobin measurement device. We also evaluate and discuss the effect of using different kinds of lighting sources.

Noninvasive hemoglobin measurement using dynamic spectrum.

Spectroscopy methods for noninvasive hemoglobin (Hgb) measurement are interfered by individual difference and particular weak signal. In order to address these problems, we have put forward a series of improvement methods based on dynamic spectrum (DS), including instrument design, spectrum extraction algorithm, and modeling approach. The instrument adopts light sources composed of eight laser diodes with the wavelength range from 600 nm to 1100 nm and records photoplethysmography signals at eight wavelengths synchronously. In order to simplify the optical design, we modulate the light sources with orthogonal square waves and design the corresponding demodulation algorithm, instead of adopting a beam-splitting system. A newly designed algorithm named difference accumulation has been proved to be effective in improving the accuracy of dynamic spectrum extraction. 220 subjects are involved in the clinical experiment. An extreme learning machine calibration model between the DS data and the Hgb levels is established. Correlation coefficient and root-mean-square error of prediction sets are 0.8645 and 8.48 g/l, respectively. The results indicate that the Hgb level can be derived by this approach noninvasively with acceptable precision and accuracy. It is expected to achieve a clinic application in the future.