Apparatus, systems, and methods for rapid cancer detection

Abstract

Presented herein are systems, methods, and apparatus that analyze molecular imprints for detecting cancerous cells. Embodiments of the present disclosure include systems, methods, and apparatus that analyze metabolic imprints of cells for cancer detection. In certain embodiments, the methods/systems comprise extracting thermal and thermodynamic quantities and properties from the molecular imprints. The thermal/thermodynamic quantities and/or further-processed quantities can be mapped on a universal cancer diagnostic scale for disease stratification, thereby providing/determining a normality status of the subject cells.

What is claimed is: 
     
         1 . A method for detecting one or more types of cancers (e.g., deciding a normality status, e.g., subtle cases of cancers) from metabolic imprints of cells, the method comprising:
 (i) accessing an image in a database;   (e.g., selecting an area to be analyzed in the image)   (ii) identifying, by a processor of a computing device, one or more cells in the image;   (iii) segmenting, by the processor, each of the one more cells into a nucleus area and a cellular area;   (iv) for each of the one or more cells, extracting, by the processor, an information surface value associated with a nuclear contrast feature (e.g., temperature difference between the nucleus area and the cellular area, e.g., contrast difference between the nucleus area and the cellular area), and a nuclear area feature (e.g., a ratio of a nucleus area to a nuclear volume projection)   (e.g., wherein the nuclear contrast feature and/or the nuclear area feature are obtained at a wavelength corresponding to a maximal radiant power of the naturally emitted IR from one or more cells),   (e.g., wherein the nuclear contrast feature is an averaged value of one or more nuclear contrast features obtained at one or more wavelengths of naturally emitted IR from one or more cells),   (e.g., wherein the nuclear area feature is an averaged value of one or more nuclear area features obtained at one or more wavelengths of the naturally emitted IR from one or more cells);   (v) calculating, by the processor, a specificity index, a log thermal capacity and a diagnostic score from the information surface values; and   (vi) determining, by the processor, a normality status for at least a portion of the image by mapping the diagnostic score on a reference scale (e.g., identifying one or more portions of the cells in the image that are cancerous) (e.g., identifying one or more stages of cancer for one or more portions of the cells in the image).   
     
     
         2 . The method of  claim 1 , wherein the specificity index is determined by
 (1) calculating each of an ensemble average of local specific heats for each of subpopulations of cells (e.g., wherein the subpopulations of cells are decided by the information surface values) (e.g., each of the subpopulations of cells corresponds to different cell cycle stages and/or different intrinsic cell cycle time), and   (2) integrating the ensemble average of local specific heats over cell cycle stages or intrinsic cell cycle time.   
     
     
         3 . The method of  claims 1 - 2 , wherein the log thermal capacity is calculated by integrating logarithm of local specific heats for each of subpopulations of cells. 
     
     
         4 . The method of  claims 1 - 3 , wherein the diagnostic score is 
       
         
           
             
               
                 
                   
                     Specificity 
                      
                     
                         
                     
                      
                     Index 
                   
                   
                     shape 
                      
                     
                         
                     
                      
                     feature 
                   
                 
                  
                 
                     
                 
                  
                 or 
                  
                 
                     
                 
                  
                 
                   
                     
                       ∏ 
                       i 
                     
                      
                     
                       Shape 
                        
                       
                           
                       
                        
                       
                         feature 
                         i 
                       
                       × 
                       Log 
                        
                       
                           
                       
                        
                       Thermal 
                        
                       
                           
                       
                        
                       Capacity 
                     
                   
                   
                     Specificity 
                      
                     
                         
                     
                      
                     Index 
                   
                 
               
               , 
             
           
         
         wherein the shape feature represents a shape of a piecewise curve, 
         wherein the piecewise curve is a curve of intrinsic cell cycle time vs the specificity index, or a curve of the intrinsic cell cycle time vs the log thermal capacity, 
         wherein the shape feature is selected from the group consisting of a number of subsections in the piecewise curve, a value of the intrinsic cell cycle time at one or more junctions in the piecewise curve, a curvature at the junction of the subsections the piecewise curve and combinations thereof, and 
         wherein i is an integer and i>0. 
       
     
     
         5 . The method of  claim 1 - 4 , further comprising one or more of (vii) to (xi) as follows:
 (vii) objectively auto-detecting, by the processor, one or more of (i) to (iii) as follows: (i) cancer rare subpopulations, (ii) subtle mimics, and/or (iii) look alike cases (that are normally difficult to diagnose by human eye and thus an extremely hard goal to be achieved by machine learning and AI techniques requiring lots of training data where the training data, to start with is provided by human experts);   (viii) following step (vii), extracting, by the processor, prognostic features comprising (including but not limited to) differentiation status and grade of the disease;   (ix) determining, by the processor, progression of disease and its possible regression after therapy/drug treatment from evolution of the diagnostic features and indices, such as the specificity index, from the (temporally) longitudinal data;   (x) determining, by the processor, in real time, efficacy of cancer drug treatment from (temporally) longitudinal data; and   (xi) detecting, by the processor, drug resistant subpopulations (that are often left unnoticed and cause the disease to recur).   
     
     
         6 . The method of  claims 1 - 5 , wherein the image is produced from a long wavelength infrared radiation (IR) (e.g., naturally emitted IR from one or more cells, e.g., having a wavelength of 3 μm to 20 μm, e.g., having a wavelength of 8 μm to 14 μm) detector. 
     
     
         7 . The method of  claim 6 , wherein the long wavelength infrared radiation detector is operated at room temperature (e.g., 15° C. to 25° C.). 
     
     
         8 . The method of  claim 6 , wherein the long wavelength infrared radiation detector is operated at physiological temperature (e.g., 35° C. to 38° C.). 
     
     
         9 . The method of  claim 6 , wherein the long wavelength infrared radiation detector is operated at a temperature of below 15° (e.g., below 10° C., below 5° C.). 
     
     
         10 . The method of  claim 6 , wherein the long wavelength infrared radiation detector is operated at a temperature substantially equal to a temperature of liquid nitrogen (e.g., −196° C.) 
     
     
         11 . The method of  claims 1 - 5 , wherein the image is an H&E stained image. 
     
     
         12 . The method of  claims 1 - 11 , wherein the one or more types of cancers are rare subpopulations, subtle mimics or look-alike cases (e.g., wherein the one or more types of cancers are difficult to be diagnosed by human eye and/or require machine learning). 
     
     
         13 . The method of  claims 1 - 12 , comprising refining the normality status using a pre-trained machine learning technique. 
     
     
         14 . The method of  claims 1 - 13 , comprising further determining, by the processor, a prognostic feature (e.g., a disease state or grade) based at least in part on the specificity index. 
     
     
         15 . The method of  claims 1 - 14 , wherein the method further comprising:
 selecting an area to be analyzed in the image, wherein the area comprises at least two adjacent sections;   repeating the steps of (i)-(v) for each of the at least two adjacent sections; and   deciding if the area comprises a cancer boundary by comparing normality statuses of the two adjacent sections (e.g., wherein two adjacent sections of the cancer boundary have different stages of cancer).   
     
     
         16 . The method of  claims 1 - 14 , wherein the method further comprises:
 providing one or more therapeutic treatments to a subject or a sample (e.g., cell culture from a subject);   repeating steps (i)-(v); and   comparing the normality status before the one or more therapeutic treatments and the normality status after the one or more therapeutic treatments.   
     
     
         17 . The method of  claim 16 , wherein the steps of repeating and comparing are performed periodically to monitor an effect of the one or more therapeutic treatments. 
     
     
         18 . The method of  claims 1 - 17 , wherein the method determines the normality status substantially in real-time (e.g., using a field-programmable gate array (FPGA)). 
     
     
         19 . A system for detecting one or more types of cancers (e.g., deciding a normality status) from metabolic imprints of cells, the system comprising:
 a processor;   a memory having instructions stored thereon, wherein the instructions, when executed by the processor, cause the processor to perform operations comprising:
 (i) accessing an image in a database; 
 (e.g., selecting an area to be analyzed in the image) 
 (ii) identifying one or more cells in the image; 
 (iii) segmenting each of the one more cells into a nucleus area and a cellular area; 
 (iv) for each of the one or more cells, extract an information surface value associated with a nuclear contrast feature (e.g., temperature difference between the nucleus area and the cellular area, e.g., contrast difference between the nucleus area and the cellular area), and a nuclear area feature (e.g., a ratio of a nucleus area to a nuclear volume projection) 
 (e.g., wherein the nuclear contrast feature and/or the nuclear area feature are obtained at a wavelength corresponding to a maximal radiant power of the naturally emitted IR from one or more cells), 
 (e.g., wherein the nuclear contrast feature is an averaged value of one or more nuclear contrast features obtained at one or more wavelengths of naturally emitted IR from one or more cells), 
 (e.g., wherein the nuclear area feature is an averaged value of one or more nuclear area features obtained at one or more wavelengths of the naturally emitted IR from one or more cells); 
 (v) calculating a specificity index, a log thermal capacity and a diagnostic score from the information surface values from the information surface values; 
 (vi) determining a normality status for at least a portion of the image by mapping the specificity index on a reference scale (e.g., identify one or more portions of the cells in the image that are cancerous, e.g., identify one or more stages of cancer for one or more portions of the cells in the image). 
   
     
     
         20 . The system of  claim 19 , wherein the specificity index is determined by
 (1) calculating each of an ensemble average of local specific heats for each of subpopulations of cells (e.g., wherein the subpopulations of cells are decided by the information surface values) (e.g., each of the subpopulations of cells corresponds to different cell cycle stages and/or different intrinsic cell cycle time), and   (2) integrating the ensemble average of local specific heats over cell cycle stages or intrinsic cell cycle time.   
     
     
         21 . The system of  claims 19 - 20 , wherein the log thermal capacity is calculated by integrating logarithm of local specific heats for each of subpopulations of cells. 
     
     
         22 . The system of  claims 19 - 21 , wherein the diagnostic score is 
       
         
           
             
               
                 
                   
                     Specificity 
                      
                     
                         
                     
                      
                     Index 
                   
                   
                     shape 
                      
                     
                         
                     
                      
                     feature 
                   
                 
                  
                 
                     
                 
                  
                 or 
                  
                 
                     
                 
                  
                 
                   
                     
                       ∏ 
                       i 
                     
                      
                     
                       Shape 
                        
                       
                           
                       
                        
                       
                         feature 
                         i 
                       
                       × 
                       Log 
                        
                       
                           
                       
                        
                       Thermal 
                        
                       
                           
                       
                        
                       Capacity 
                     
                   
                   
                     Specificity 
                      
                     
                         
                     
                      
                     Index 
                   
                 
               
               , 
             
           
         
         wherein the shape feature represents a shape of a piecewise curve, 
         wherein the piecewise curve is a curve of intrinsic cell cycle time vs the specificity index, or a curve of the intrinsic cell cycle time vs the log thermal capacity, 
         wherein the shape feature is selected from the group consisting of a number of subsections in the piecewise curve, a value of the intrinsic cell cycle time at one or more junctions in the piecewise curve, a curvature at the junction of the subsections the piecewise curve and combinations thereof, and 
         wherein i is an integer and i>0. 
       
     
     
         23 . The system of  claim 19 - 22 , wherein the instructions, when executed by the processor, cause the processor to perform operations further comprising one or more of (vii) to (xi) as follows:
 (vii) objectively auto-detecting, by the processor, one or more of (i) to (iii) as follows: (i) cancer rare subpopulations, (ii) subtle mimics, and/or (iii) look alike cases (that are normally difficult to diagnose by human eye and thus an extremely hard goal to be achieved by machine learning and AI techniques requiring lots of training data where the training data, to start with is provided by human experts);   (viii) following step (vii), extracting, by the processor, prognostic features comprising (including but not limited to) differentiation status and grade of the disease;   (ix) determining, by the processor, progression of disease and its possible regression after therapy/drug treatment from evolution of the diagnostic features and indices, such as the specificity index, from the (temporally) longitudinal data;   (x) determining, by the processor, in real time, efficacy of cancer drug treatment from (temporally) longitudinal data; and   (xi) detecting, by the processor, drug resistant subpopulations (that are often left unnoticed and cause the disease to recur).   
     
     
         24 . The system of  claims 19 - 23 , the system comprises a long wavelength infrared radiation (e.g., naturally emitted IR from one or more cells, e.g., having a wavelength of 6 μm to 12 μm) detector (e.g., wherein the detector does not need a light source). 
     
     
         25 . The system of  claim 24 , wherein the image is produced from the long wavelength infrared radiation (IR) detector. 
     
     
         26 . The system of  claims 19 - 23 , wherein the image is an H&E stained image. 
     
     
         27 . The system of  claims 19 - 26 , wherein the one or more types of cancers are rare subpopulations, subtle mimics or look-alike cases (e.g., wherein the one or more types of cancers are difficult to be diagnosed by human eye and/or require machine learning). 
     
     
         28 . The system of  claims 19 - 27 , wherein the instructions, when executed by the processor, cause the processor to perform operations further comprising refining the normality status using a pre-trained machine learning technique. 
     
     
         29 . The system of  claims 19 - 28 , wherein the instructions, when executed by the processor, cause the processor to perform operations further comprising further determining a disease state based at least in part on the specificity index. 
     
     
         30 . The system of  claim 19 - 29 , wherein the instructions, when executed by the processor, cause the processor to perform operations further comprising:
 selecting an area to be analyzed in the image, wherein the area comprises at least two adjacent sections;   repeating steps (ii)-(v) for each of the at least two adjacent sections; and   deciding if the area comprises a cancer boundary by comparing normality statuses of the two adjacent sections (e.g., wherein two adjacent sections of the cancer boundary have different stages of cancer).   
     
     
         31 . The system of  claims 19 - 29 , wherein the instructions, when executed by the processor, cause the processor to perform operations further comprising:
 providing one or more therapeutic treatments to a subject or a sample (e.g., cell culture from a subject);   repeating steps (i)-(v); and   comparing the normality status diagnosed before the one or more therapeutic treatments and after one or more therapeutic treatments.   
     
     
         32 . The system of  claim 31 , wherein steps repeating and comparing are performed periodically to monitor an effect of the one or more therapeutic treatments. 
     
     
         33 . The system of  claim 19 - 32 , the system determines the normality status substantially in real-time (e.g., using a field-programmable gate array (FPGA)). 
     
     
         34 . An apparatus for creating a long wavelength infrared (e.g., naturally emitted IR from one or more cells, e.g., having a wavelength of 4 μm to 20 μm) image (e.g., 1-dimensional, 2-dimensional, 3-dimensional) for detecting one or more types of cancers, the apparatus comprising:
 a spatial magnification module (e.g., having a resolution of 20 microns or less) providing an optical path; 
 an infrared radiation detector (e.g., comprising a focal-plane array, comprising a vanadium oxide surface, and/or comprising a mercury cadmium telluride surface) for detecting long wavelength infrared radiation emitted from a subject (e.g., measuring photon flux); 
 a signal processing module; and 
 a memory, the signal processing module operable with the memory to record a signal corresponding to infrared radiation detected by the infrared radiation detector and to cause rendering of an image (e.g., 1D, 2D, or 3D image) corresponding to the signal 
 (e.g., wherein the apparatus is operable to measure photon flux from the subject) 
 (e.g., wherein the apparatus is operable without an external light source) 
 (e.g., wherein the apparatus further comprises a heat source for enhancing the infrared radiation emitted from the subject) 
 (e.g., wherein the apparatus is operable to produce the long wavelength infrared radiation image in any one of the preceding claims). 
 
     
     
         35 . The apparatus of  claim 34 , wherein the signal processing module comprises a field-programmable gate array (FPGA). 
     
     
         36 . The apparatus of  claims 34 - 35 , further comprising an attachment locating the infrared radiation detector nearby (e.g., within 10 cm, 5 cm or 1 cm) an area (e.g., internal sites of human) to be diagnosed. 
     
     
         37 . The apparatus of  claims 34 - 36 , wherein the attachment is a biopsy needle. 
     
     
         38 . The apparatus of  claim 34 - 37 , wherein the spatial magnification module comprises a first biconvex lens having a first focal length (f1), a second biconvex lens having a second focal length (f2), and a third biconvex lens having a third focal length (f3)
 (e.g., wherein a distance between the object and the first biconvex lens is no less than f1 (e.g., f1 is in range from 1 mm to 5 mm), so that the first biconvex lens generates a first magnified image (e.g., a real inverted magnified image) of the subject)   (e.g., wherein a distance between the first biconvex lens and the first magnified image is no less than 2*f1)   (e.g., wherein a distance between the first magnified image and the second biconvex lens is less than f2)   (e.g., wherein the second biconvex lens re-magnifies the first magnified image, thereby generating a second magnified image)   (e.g., wherein the third biconvex lens focuses the second magnified image on the infrared radiation detector)   (e.g., wherein the spatial magnification module further comprises an aperture for filtering the naturally emitted IR from one or more cells).   
     
     
         39 . The apparatus of  claim 34 - 37 , wherein the spatial magnification module comprises a convex mirror, a concave mirror, and a flat mirror
 (e.g., wherein the convex mirror and the concave mirror are substantially concentric)   (e.g., wherein a reflective surface of the convex mirror faces a reflective surface of the concave mirror)   (e.g., wherein the concave mirror comprises on opening, so that a reflected radiation from the convex mirror passes through the opening) (e.g., wherein the opening is located on the optical path).   
     
     
         40 . A handheld diagnostic imaging device for performing one or more of the methods of  claims 1 - 18 , the device comprising:
 relay optics (e.g., transmission and/or reflective optics) for transmitting a signal from a biological sample (e.g., in vivo, in vitro, or ex vivo) to a detector;   the detector (e.g., a 2-D detector);   a signal processing module (e.g., FPGA running algorithm) and a memory for processing a signal received from the detector; and   optionally, a display for providing a diagnostic result determined by the signal processing module.