The invention relates to a novel method for
breast cancer prognosis evaluation based on collagen distribution in a
tumor microenvironment. The method comprises the steps of dividing the development ofcollagen fibers in the
tumor microenvironment into three stages and eight collagen morphological distribution characteristics; performing multi-
photon scanning imaging on a dewaxed
breast cancer paraffin embedded tissue slice to obtain second
harmonic images, namely,
collagen fiber distribution images; classifying the
collagen fiber distribution in each second
harmonic image to obtain corresponding scores of the eight collagen distribution characteristics of the whole slice; dividing the
score of each collagen morphological
distribution characteristic by the total
score of the eight collagendistribution characteristics to calculate the percentage of each collagen characteristic; performing
cross validation by adopting
ridge regression analysis, and calculating to obtain the weight of each collagen characteristic; and performing weighted sum on the percentages of the eight collagen morphological distribution characteristics and the corresponding eight weights, and calculating the final collagen distribution
score of each patient, wherein the final collagen distribution score is used for judging the probability that the patient has five-year
disease-free survival time. According tothe method, the postoperative risk condition of the patient can be simply, accurately and rapidly evaluated.