Smartphone supported backlight illumination and image acquisition for microfluidic-based point-of-care testing

: A smartphone-based image analysis system is advantageous for point-of-care testing applications. However, the processes of observation and image recording rely heavily on an external attachment that includes additional light sources. Moreover, microfluidic point-of-care devices are highly miniaturized, and can be clearly observed only under magnification. To address these issues, the present work proposes a novel imaging box for converting the built-in light source of a smartphone into uniform backlight illumination to avoid interference arising from reflections. A multi-piece orthoscopic lens is embedded in the imaging box to enable the imaging of micro-sized samples. As such, the colorimetric signal of a microchannel with a width as small as 25 µm can be faithfully recorded. Protein concentration quantification based on the bicinchoninic acid assay method was demonstrated with the proposed smartphone/imaging box system from an analysis of colorimetric signals. In addition, a microfluidic chip for conducting ABO blood typing was fabricated, and the microscopic imaging of induced blood coagulation can be clearly observed in a 3 µL sample using the proposed system. These results highlight the potential for adopting smartphone-based analysis systems in point-of-care testing applications.

the colorimetric content of color digital images can be numerically analyzed by retrieving pixel intensities, and several color models, such as red-green-blue (RGB) and the huesaturation-value [13][14][15], have been proposed to quantitatively compare different color signals. As such, the built-in camera of a smartphone can effectively capture color and structural information on microfluidic devices, and the integration of a smartphone camera and a microfluidic chip effectively reduces the complexity and cost of a POCT system [16,17]. However, the direct use of smartphones for imaging suffers from non-uniform illumination caused by the light source and the surrounding environment [17]. Meanwhile, the relative positioning between the sample and the smartphone also introduces bias into colorimetric analyses. For this concern, Jung et al. developed a miniaturized attachment to provide stable illumination conditions, as well as to provide a uniform positioning between the sample and the smartphone [18]. Two types of external lighting have been adopted in imaging boxes. The first type arranges an external light source above the sample [1,5,9,17]. However, this approach results in image exposure problems associated with light reflection from the sample. The second type supplies a backlighting source that illuminates the sample from the bottom, which can eliminate exposure problems caused by light reflection [2,6,7,10,15]. An additional advantage of backlighting is that it enables capture of the transmission signals of light passing through colorimetric samples. Venkatesh et al. designed a detection device containing an array of light emitting diodes (LEDs) to provide uniform illumination of samples, and enzyme-linked immunosorbent assay (ELISA) was conducted based on the colorimetric information of microplate images retrieved from the smartphone. However, past efforts at developing smartphone-based POCT systems have uniformly neglected the convenience of the light source installed on the smartphone itself. A miniaturized device that can be docked on a smartphone, and which utilizes the light source of the smartphone itself to provide backlighting illumination would be most favorable owing to its greater simplicity and portability. In addition, the micro-size structures of microfluidic devices cannot be clearly imaged by the built-in camera of a smartphone. Therefore, smartphone-based POCT systems must also provide for enhanced magnification of device structures.
To address these issues, we designed a miniature imaging unit that can be docked onto a smartphone. We adopted an optical configuration consisting of an optical fiber, a mirror, and a light guide plate (LGP) to convert the built-in smartphone light source into a planar backlighting source. In addition, a set of orthoscopic lens was assembled to facilitate the clear imaging of small-sized devices. The two functional units were assembled in a threedimensional (3D) printed epoxy box. The viability of the proposed smartphone/imaging system was demonstrated by conducting on-chip protein concentration measurements based on the bicinchoninic acid (BCA) method and ABO blood typing. The reduced sample volume employed in the system enabled the successful measurement of protein concentrations and the conduct of ABO typing without special equipment. These results highlight the potential of for adopting smartphone-based analysis systems in POCT applications.

Design and fabrication of the imaging box and different microfluidic chips
The imaging box consists of an illumination system, a magnification module, and a mechanical adaptor. The top plate includes two circular apertures for docking the camera lens and the LED light source of the smartphone, which were positioned according to the geometrical layout of the camera. The present study employed an Apple iPhone 6s smartphone. Therefore, the camera lens and light source apertures were 8 mm and 5 mm in diameter, respectively, as shown in Fig. 1(A). The microfluidic stage was designed at the bottom of the imaging box under the imaging system, and a sliding gate on the side of the box allows for sample input and output. The individual systems are discussed in detail as follows.
Illuminatio re-direct the composed of p 15 mm. To av black tape, an was connecte to redirect th enter an LGP to produce ev Fig. 2(D). Th employed in t total internal r produced by bottom of the mounted on th Magnifica groups with a single plano-c reduce lens d alleviate chro lenses were a length of 13 according to t in Fig. 2 under the came ation system w = 10.5 mm. Th nted convex-co n of the magn rtion in the ce 3D-printed hol ertically assem Fig. 1(B). The tion factor form nce of the ma h most dispos and paper-base martphone-based ematic of the intern atic of the convert he guide light plat e smartphone via a n of this syste nnected to the nserting the ph orresponding a shown in Fig. illumination. T d a diameter o of the optical f e aperture. The 45° relative to ce ( Fig. 1(C) ), and is e adapter rly aligns x, which greatly simplifies the image capture process. Hence, the smartphone can be easily connected to the imaging box through the smartphone cover-like adapter. The imaging box and smartphone cover-like adapter were printed using a MoonRay 3D printer (SprintRay Inc., USA). The lenses, optical fiber, mirror, and LGP were then assembled in the imaging box.
Two types of miniaturized devices were tested to demonstrate the image capacity of the box. For smartphone imaging, the light source was set in its continuous operation mode, the camera was focused on the target chip, and then the image was captured according to standard procedure. First, Formvar/carbon film coated grids used in transmission electron microscopy (TEM) analysis were imaged to test the magnification effect and imaging quality of the imaging box. The mesh size of the carbon coated grids was 230, while the bar width was 25 μm, the pitch was 110 μm, the grids number of the mesh was 500 and the hole size was 85 μm. In order to compare the quality of photographs taken with and without the image box, Image Pro plus (version IPP6.0, MediaCybernetics, USA) was applied to count the number of grids. Specifically, particle counting and measurement function was applied to analyze the image to automatically extract and count the grids. Second, a PDMS device was fabricated following standard lithography and replication techniques [19]. Structured PDMS was replicated from a photoresist mold, and assembled with an adhesive film to form the microchannels. A color solution composed of food additive dye and water in a 1:20 v/v mixture was injected into the microchannels. To compare the quality of photographs taken with and without the image box, Plot Profile function of ImageJ software (NIH, USA) was used to analyze the grayscale along the edge of the channel.

Testing the colorimetric analysis abilities of the imaging box
As discussed, the potential of the proposed imaging box in POCT applications was demonstrated by conducting colorimetric-based protein detection and ABO antigen-based blood typing using microdevices. These analyses are discussed in detail as follows.
Bicinchoninic acid (BCA) protein detection: BCA protein detection is a standard colorimetric assay conducted in laboratories. The working principle of the assay lies in that monovalent copper ions (Cu + ) interact with a BCA reagent to form a purple reactive complex, which exhibits a strong absorbance at 562 nm. Because the peptide bonds in protein molecules reduce Cu 2+ ions from copper (II) sulfate to Cu + , the amount of Cu 2+ reduced is proportional to the amount of protein present in the solution. Thus, the colorimetric changes (absorbance value) would reflect the protein concentration in a sample. For conducting BCA protein detection, different concentrations of bovine serum albumin (BSA) solution (0, 0.05, 0.1, 0.2, 0.3, 0.4, and 0.5 mg/ml) was added into the BCA reagent according to instructions (BCA protein kit, Beyotime Biotechnology, China), and 20 μL samples were incubated at 37°C for 45 min in compartments with a diameter of 3 mm and a depth of 1.5 mm that were fabricated by 3D printing. The colorimetric variation was imaged by the smartphone/imaging box system, and the captured images were quantitatively analyzed using ImageJ software according to the G channel intensities based on the RGB model in the detection zone. Next, the pixel intensity of G channel was direct read by the smartphone APP color name (Softporek BG Ltd.). In addition, the same BCA solution samples were analyzed using a spectrophotometric microplate reader (ELx800TM, Gene Company) at 570 nm. The signal (intensity of G channel or absorbance) to protein concentration curve was plotted. All experiments were repeated for three times.
ABO blood typing: The ABO blood type system is classified according to whether specific antigens (agglutinogens) A and B exist on the surface of red blood cells [20]. Blood is then divided into four types: A, B, AB, and O, according to the relative distributions of A and B. First, the flexible thin film device shown in Fig. 5(A) was fabricated using shaped Parafilm and thermal lamination with two transparent PET films to form micro-channels [21]. The chip is a tridentate structure containing three reaction zones, each of which is a circle with a d sub-channel i layer, which antibody, anti zones and inc sheet. Blood University H diluted blood channel and r corresponding Images of th system. In the considered. The image imaging a car structure of t function of th would compr coated grid is system. The t in Fig. 3(A) Fig. 3(B) ut the assist of how the image the profile fun omised pixel q ness profile lin Figure 3(A) an ation and capt the proposed e. The illumina the orthoscopi size. The all-in A) Carbon grid ca mage box ; particl martphone zoom in ation at the right odel.

hone based imag
Ref Backligh illuminatio [1] Yes [3] Yes [5] Yes [17] Yes [13] Yes [18] Yes d colorimetric fferent BSA co results are pre bserved with in en with the im ts of image an Fig. 4 with the al change n from 0 omputernsity) for he results ess of fit detection limit of 0.041 signal divide detection sam color name ( protein conce range of 0-0.5 and the calcu obtained by th Fig. 4(D)

On-chip
An image of Fig. 5(A). We coagulation b smartphone/im detection zone blood typing a

Conclusio
The present w could take fu miniaturized the LED ligh system consis lens was asse region of mic colorimetric-b the colorimet clearly record results highlig