Back to results
US 9,739,783 B1Granted
Convolutional neural networks for cancer diagnosis
AssigneeAnixa Diagnostics Corporation
InventorsAmit Kumar, John Roop
Life of this patent
Estimating the term…
Abstract
The present disclosure provides devices and systems for diagnosing and characterizing cancer in a subject. Devices include microfilters and microfiltration systems useful for the isolation and characterization of cells from a subject.
Claims28
A method of determining the presence or absence of a target cell in a biological sample from a subject, the method comprising:
- a) passing the biological sample through a microfilter comprising a metal layer and a plurality of pores extending through the metal layer;
- wherein a length of one or more of the plurality of pores is smaller than a length of the target cell, and the metal layer comprises nickel;
- b) if the biological sample comprises a cell of interest, retaining the cell of interest on a surface of the metal layer;
- c) applying to the surface of the metal layer a first probe comprising a molecule specific for a biomarker of the target cell;
- and
- d) if the cell of interest is retained on the surface of the metal layer, determining if the first probe is bound to the cell of interest;
- wherein if the first probe is bound to the cell of interest, the cell of interest is the target cell and the target cell is present in the biological sample from the subject, wherein the first probe comprises a first fluorophore;
- and
- e) determining if the first probe is bound to the cell of interest using a computer-implemented method comprising:
- (i) transmitting, by a computer, a surface scanning instruction to a multichannel fluorescence acquisition device, wherein the microfilter is positioned on a stage of the multichannel fluorescence acquisition device;
- the surface scanning instruction comprising one or more commands for the multichannel fluorescence acquisition device to determine the z-axis height of the microfilter surface;
- (ii) transmitting, by a computer, an image capture instruction to the multichannel fluorescence acquisition device, the image capture instruction comprising one or more commands for the multichannel fluorescence acquisition device to acquire a plurality of digital images of the microfilter using a channel corresponding to the first fluorophore,
- wherein the plurality of digital images are acquired at each stage position in a z-stack having a focus plan centered on the microfilter surface height determined in step (i);
- (iii) generating, by a computer, an in-focus image by stitching together in-focus portions of each of multiple images of the plurality of digital images;
- (iv) selecting, by a computer, an in-focus image that comprises the cell of interest;
- and
- (v) identifying, by a computer, an area of interest comprising the cell of interest in the in-focus image.
14 more elements
A method of diagnosing cancer in a subject, the method comprising:
- a) passing a biological sample from the subject through a microfilter comprising a metal layer and a plurality of pores extending through the metal layer;
- wherein a length of one or more of the plurality of pores is smaller than a length of a target cell, and the metal layer has an autofluorescence less than a membrane filter comprising parylene-C in the FITC, TRITC or Cy5 fluorescent channel;
- b) if the biological sample comprises the target cell, separating the target cell from the biological sample on a surface of the metal layer, and (i) retaining the target cell on the surface of the metal layer or (ii) collecting the target cell from the surface of the metal layer;
- c) applying a probe comprising a molecule specific for a biomarker of the target cell to either (i) the surface of the metal layer or (ii) material collected from the surface of the metal layer;
- wherein if the biological sample comprises the target cell, the target cell is present on either the surface of the metal layer or the material collected from the surface of the metal layer, and the probe is bound to the target cell, wherein the probe comprises a fluorophore;
- and
- d) determining the presence or absence of the target cell on either (i) the surface of the metal layer or (ii) the material collected from the surface of the metal layer, by detecting the presence, absence, intensity, location, or combination thereof of a signal from the applied probe using a computer-implemented method comprising:
- I. transmitting, by a computer, a surface scanning instruction to a multichannel fluorescence acquisition device, wherein the microfilter is positioned on a stage of the multichannel fluorescence acquisition device;
- the surface scanning instruction comprising one or more commands for the multichannel fluorescence acquisition device to determine the z-axis height of the microfilter surface;
- II. transmitting, by a computer, an image capture instruction to the multichannel fluorescence acquisition device, the image capture instruction comprising one or more commands for the multichannel fluorescence acquisition device to acquire a plurality of digital images of the microfilter using a channel corresponding to the fluorophore, wherein the plurality of digital images are acquired at each stage position in a z-stack having a focus plan centered on the microfilter surface height determined in step (I);
- III. generating, by a computer, an in-focus image by stitching together in-focus portions of each of multiple images of the plurality of digital images;
- IV. selecting, by a computer, an in-focus image that comprises the cell of interest;
- and
- V. identifying, by a computer, an area of interest comprising the cell of interest in the in-focus image;
- wherein the presence of the target cell is indicative of cancer in the subject.
12 more elements
Dependencies are inferred from the claim text.