Image acquisition methods for simultaneously detecting genetic rearrangement and nuclear morphology

A method of determining a cancer sub-type in a single cell or single sample derived from a subject diagnosed with cancer, the method comprising:

• a. contacting the single cell or the single sample with a fluorescent nuclear stain for making visible sizes of nuclei, shapes of nuclei, and amount of stain contained in nuclei, and subjecting the single cell or single sample to fluorescence in situ hybridization (FISH) for detecting the genetic rearrangement marker using labeled nucleic acids;
• b. performing a single acquisition of multivariate image data to acquire a raw spectral data cube for each of a predetermined set of wavelengths ranging from between about 400 nm to about 800 nm in a single field of view of the single cell or single sample using a calibrated quantitative light source and a calibrated quantitative spectral detection system, wherein the calibrated quantitative light source comprises a with a light source stabilized over the predetermined set of wavelengths and which is capable of repeating illumination at a plane of the single cell or the single sample with less than 1% variation in absolute illumination level over the predetermined set of wavelengths such that variability in brightness levels reflects an actual stain distribution within the single cell or in the single sample, wherein the raw spectral data cube comprises intensities for each wavelength at each pixel;
• c. using linear unmixing to deconvolve the raw spectral data cube into specific wavelength channels representing a relative distribution of signal at each pixel, generating individual images for each channel of interest, then generating an image representing nuclear morphology and an image representing the genetic rearrangement marker;
• d. calculating for each nucleus except non-glandular nuclei or non-cancer nuclei one or more of:
• (i) nuclear roundness;
• (ii) relative chromatin content;
• (iii) intranuclear chromatin distribution;
• (iv) nuclear size;
• (v) presence of macronulceoli;
• (vi) nuclear shape;
• (vii) intranuclear variability of chromatin labeling;
• and (viii) status of the genetic rearrangement marker;
• and
• e. determining a cancer sub-type by comparing the one or more of:
• (i) nuclear roundness;
• (ii) relative chromatin content;
• (iii) intranuclear chromatin distribution;
• (iv) nuclear size;
• (v) presence of macronulceoli;
• (vi) nuclear shape;
• (vii) intranuclear variability of chromatin labeling;
• and (viii) status of the genetic rearrangement marker of each cell to data of (i) nuclear roundness;
• (ii) relative chromatin content;
• (iii) intranuclear chromatin distribution;
• (iv) nuclear size;
• (v) presence of macronulceoli;
• (vi) nuclear shape;
• (vii) intranuclear variability of chromatin labeling;
• or (viii) status of the genetic rearrangement marker from a library of reference cancer cell populations to determine a cancer sub-type.

A method of predicting a clinical outcome of a subject having cancer, the method comprising:

• a. contacting a single cell or a single sample derived from the subject with (i) a nuclear stain for making visible sizes of nuclei, shapes of nuclei, and an amount of stain contained in nuclei, and (ii) one or more labeled nucleic acids specific to a genetic rearrangement marker;
• b. performing a single acquisition of multivariate image data to acquire a raw spectral data cube for each of a predetermined set of wavelengths ranging from about 400 nm to about 800 nm in a single field of view of the single cell or the single sample using a calibrated quantitative light source and a calibrated quantitative spectral detection system, wherein the raw spectral data cube comprises intensities for each wavelength at each pixel;
• c. using linear unmixing to deconvolve the raw spectral data cube into specific wavelength channels representing a relative distribution of signal at each pixel, generating individual images for each channel of interest, then generating an image representing nuclear morphology and an image representing the genetic rearrangement marker;
• d. calculating for each nucleus except non-glandular nuclei or non-cancer nuclei one or more of (i) nuclear roundness;
• (ii) relative chromatin content;
• and (iii) status of the genetic rearrangement marker;
• e. determining a cancer sub-type comparing the nuclear roundness, relative chromatin content, and status of the genetic rearrangement marker of each cell to nuclear roundness, relative chromatin content, and genetic rearrangement marker data from a library of reference cancer cell populations;
• f. obtaining a Gleason score for the single cell or single sample derived from the subject;
• and
• g. predicting an outcome based on the determined cancer sub-type and obtained Gleason score.