(J-362) pHast Cam: Analysis of Paper-based pH Blood Sensors via Smartphone as Birth Asphyxia Screening Tool

Birth asphyxia occurs when a lack of oxygen or blood flow to a baby’s brain and other organs occurs, usually secondary to a failure to breathe soon after birth. Twenty-three percent of neonatal deaths globally are caused by birth asphyxia, which represents nearly 1 million preventable deaths annually [1]. Birth asphyxia causes brain and neurological damage, called hypoxic ischemic encephalopathy (HIE), and early detection of babies with asphyxia in the first 6 hours can lead to therapeutic interventions, such as therapeutic hypothermia, that can prevent death or severe morbidity [2]. In high-income settings, HIE is typically screened by measuring blood pH. Unfortunately, the diagnostic equipment required to assess blood pH is impractical for low-resourced settings globally because it is costly (i.e. $20/test and $5,000 for equipment) and requires technical expertise to operate. We aim to develop a cost-effective device for HIE screening for use in low-resourced settings that is a user-friendly, paper-based blood pH sensor. This project requires an analysis of pH-sensitive paper to screen for HIE, which has a clinical range of 7.0-7.2 in less than 0.1 pH units to distinguish thresholds for high HIE risk (pH < 7.1), medium risk (pH 7.1 – 7.2), and low risk (pH > 7.2) [3]-[5]. The device combines an inexpensive pH sensitive dye, a smartphone camera, and a 3D printed integrating sphere that controls the imaging environment to quickly identify acidosis that results from HIE.

Materials and Methods:

Our paper-based pH sensor incorporates a pH-sensitive dye, bromothymol blue (BTB), which has a well-characterized absorption spectra that detects subtle pH changes within the clinically-relevant pH range of 6.0 to 7.6 [6]. To distinguish between pH values between 6.8 and 7.2, we first maximize color change to aid in distinguishing 0.1 pH units. According to Beer's law, increasing the concentration of BTB dye will lead to a greater intensity of color change, thereby improving differentiation between high and low pH regions associated with HIE.
We prepare our pH sensors by coating MicroEssential Lab 235 paper strips with the solution containing BTB (~50%) and additives such as Tween20 (~3%) for stability and uniformity. Each pH sensor consists of four pH paper sites, with buffer solution in our pH range of interest deposited in 3 µL volumes and enclosed by a hydrophobic black paper mask (Figure 1).
The pH sensor is placed inside a 3D printed integrating sphere, along with an red-green-blue (RGB) LED emitter (Figure 2) . An integrating sphere is used to produce lighting homogeneity for reliable data collection in optical engineering. Our imaging environment hardware utilizes red LED light because BTB dye is most sensitive to light between 600 to 640 nm (Figure 3). A smartphone is positioned on the sphere to capture images at optimal parameters (ISO 100 and 1/10 second shutter speed). Image capture occurs at three-minute intervals, with sets of five images taken to determine average results.

Results, Conclusions, and Discussions:

We measured pH at 6.86, 7.0, 7.2, and 7.4 pH units with our paper-based sensors with BTB dye and imaging using our integrating sphere and smartphone. Our data demonstrates accuracy of 0.1 ±0.04 pH units. At this stage, we have achieved a regressive linear model that predicts buffered solution acidity from a pH of 6.86 to 7.4 (Figure 4).
Our 3D-printed integrating sphere provides a secure platform to control the camera-to-target distance and uniform lighting between images (Figure 2). The normalized digital number range over this 500x500 pixel area is 0.96 to 0.97, resulting in a consistency of ±0.5%. As our sensor targets are roughly 20x20 pixels, the uniformity of the current fixture is well within our requirements.
Our team has also optimized the parameters applied to the camera by using a fixed ISO value (100) and a fixed shutter speed (1/10s). This coupled with a fixed red luminance enhances the slope of our model (Figure 4). The purpose of altering the original ISO value and shutter speed of the camera was to capture images that collect all the relevant data required for image analysis. The digital number average is obtained by signal averaging 5 images followed by averaging the resulting 20x20 pixel region of interest. Therefore, each average is calculated from 400 pixels times 5 images (2000 pixels total). The final aspect of our imaging environment is the imaging sphere which allows for uniform images by avoiding shadows and lighting discrepancies.
In conclusion, we can distinguish between pH values that indicate low, medium, and high HIE risk using buffer solutions and our custom sensor and imaging platform. Moving forward, we plan to test whole blood samples and are investigating strategies to minimize the interaction between albumin, a protein in blood, and our dye indicator BTB. We plan to include a plasma separation membrane layer by Pall to filter red blood cells. We also plan to obtain a wider range of digital numbers and expand the dynamic range of our pH sensor by using 10-bit raw image data from the smartphone camera.

Acknowledgements (Optional):

Contributors: Dr. Ayokunle Olanrewaju, Dr. Krystle Perez, Tim Robinson, Dr. Gregory Valentine, Dr. Manuja Sharma, Dr. Eric Seibel, Eric Fan, Shivesh Ummat, and Caleb Berhow.

Funding: This work was supported by awards from M.J. Murdock Diagnostics Foundry for Traditional Research, University of Washington CoMotion Innovation Gap Fund, and University of Washington Royalty Research Fund.

References (Optional):

[1] "Diaz-Rosello JGP, Niermeyer S, et al. WHO Basic guidelines on new born resuscitation. 2012."

[2] Soll, Roger F. "Cooling for newborns with hypoxic ischemic encephalopathy." Neonatology 104.4 (2013): 260.

[3] Wiberg, Nana, et al. "Relation between umbilical cord blood pH, base deficit, lactate, 5‐minute Apgar score and development of hypoxic ischemic encephalopathy." Acta obstetricia et gynecologica Scandinavica 89.10 (2010): 1263-1269.

[4] Toh, V. C. "Early predictors of adverse outcome in term infants with post‐asphyxial hypoxic ischaemic encephalopathy." Acta Paediatrica 89.3 (2000): 343-347.

[5] Vesoulis ZA, Liao SM, Rao R, Trivedi SB, Cahill AG, Mathur AM. Re-examining the arterial cord blood gas pH screening criteria in neonatal encephalopathy. Arch Dis Child Fetal Neonatal Ed. 2018 Jul;103(4):F377-F382. doi: 10.1136/archdischild-2017-313078. Epub 2017 Sep 23.

[6] Abcam plc., “Bromothymol blue, pH indicator (CAS 76-59-5) (ab146293) | Abcam,” www.abcam.com, 2023. https://www.abcam.com/products/biochemicals/bromothymol-blue-ph-indicator-ab146293.html