Computational protocol: Quantitative diagnosis of HER2 protein expressing breast cancer by single‐particle quantum dot imaging

Similar protocols

Protocol publication

[…] To quantitatively measure the particle numbers of QDs bound to cancer tissues, analysis was carried out as follows : The 512‐pixel‐square images, filtered with 695–740 nm band‐pass filter for QDs and auto‐fluorescence of tissues or a 640–690 nm band‐pass filter for the auto‐fluorescence, were taken at an exposure time of 20 sec. Each image was converted into a JPEG file. During the conversion, the autofluorescent signal of the image filtered with the 640–690 nm band‐pass filter (640–690 nm image) was adjusted to be about 1.2‐fold greater than that filtered with the 695–740 nm band‐pass filter (695–740 nm image). After file conversion, in order to visualize only the signal of the fluorescent QDs, the JPEG image of the 640–690 nm image was subtracted from that of 695–740 nm image using Adobe Photoshop image processing software. The fluorescent intensity of the QD signal in the subtracted image was analyzed as gray values using ImageJ software ( The total fluorescent intensity in the image was defined as total‐QDs value. In the subtracted image, we also confirmed the fluorescent intensity in an area where there was an autofluorescent signal of zero. This indicates that, except for the signal of the QDs, there is no fluorescent signal in the background of the subtracted images. To precisely measure the number of the QD particles on the tissues, it is necessary to define the fluorescent intensity of a single QD. As QDs possessing the same fluorescent wavelength are uniform in size, the fluorescent intensity of QDs is proportional to the particle number. In addition, QD fluorescence is composed of fluorescent and nonfluorescent states called on‐ and off‐states. This fluorescent property results in blinking of QDs . When we measured the fluorescence of fresh QD particles after purchasing and analyzed their property, the results showed that the mean time of the off‐state during 20 sec of observation was about 4 sec and the calculated S.E.M. value was very low . If several QDs are aggregated, the mean time of the off‐state per unit time is shortened by aggregation of QDs because the on‐state and off‐state of each particle in the aggregate occurs randomly. Therefore, based on an off‐state time of 4 sec, we selected a single‐particle QD using each subtracted image and video image and measured the fluorescent intensity of the single QD particle (single‐QD value). In addition, the cell number in each image was measured using the DAPI image (300–600 cells were investigated in a patient sample). Then, the total‐QD value was divided by the single‐QD value to calculate the number of QD particles in a cell. Finally, we subtracted the value of the image labeled with QD‐conjugated human IgG (particle number/cell) from that labeled with QD‐conjugated trastuzumab, and obtained the precise mean particle number of QD‐conjugated trastuzumab bound specifically to a cancer cell. [...] We randomly selected 37 patients with breast cancer, based on the result that 18 or more patient samples were required to statistically verify the strong correlation between IHC‐QD scores and FISH scores (R ≥ 7.0). The Pearson correlation method was used to compare the results between IHC‐QDs and FISH or TTP, considered as continuous variables. All statistical analyses were performed using the SAS software, JMP Pro 11. Two‐tailed P < 0.05 was considered statistically significant. […]

Pipeline specifications

Software tools ImageJ, JMP Pro
Applications Miscellaneous, Microscopic phenotype analysis
Organisms Homo sapiens
Chemicals 3,3'-Diaminobenzidine