On 2016 Jan 29, Daniel Corcos commented:

Welch et al. draw unwarranted conclusions from the incidence curves they show. In their paper, there is no information about the incidence of prostate cancer before initiation of widespread screening. However, as PSA screening is more likely to detect metastatic cancer, since high levels of PSA are associated with metastasis (1), it is very likely that screening leads to early detection of metastatic cancer, with the expected consequence of a peak of detected incidence at initiation of screening, followed by a return to basal level. In contrast mammography screening detects only primary cancers, with no bias for metastatic cancer. The incidence of metastatic cancer should be related to the incidence of primary cancer, which has risen during the same period (2). If breast cancers were metastatic at the time of detection as they claim, there would be no benefit from early surgery (3).

I conclude that there is no difference in cancer dynamics that could be inferred from these incidence curves, and that early detection of breast cancer is still the best option in order to shorten time to surgery. However, a 30% increase of primary breast cancer after implementation of x-rays screening is worrying and suggests that x-ray screening may be carcinogenic, at least in women with hereditary DNA repair defect (4). A retrospective study comparing the percentage of DNA repair defects in metastatic breast cancer before and after initiation of widespread mammography would be useful.

1) http://www.ncbi.nlm.nih.gov/pubmed/26659430 2) http://seer.cancer.gov/ 3) http://www.ncbi.nlm.nih.gov/pubmed/15126770 4) http://www.ncbi.nlm.nih.gov/pubmed/22788243

On 2016 Jan 29, Daniel Corcos commented:

Welch et al. draw unwarranted conclusions from the incidence curves they show. In their paper, there is no information about the incidence of prostate cancer before initiation of widespread screening. However, as PSA screening is more likely to detect metastatic cancer, since high levels of PSA are associated with metastasis (1), it is very likely that screening leads to early detection of metastatic cancer, with the expected consequence of a peak of detected incidence at initiation of screening, followed by a return to basal level. In contrast mammography screening detects only primary cancers, with no bias for metastatic cancer. The incidence of metastatic cancer should be related to the incidence of primary cancer, which has risen during the same period (2). If breast cancers were metastatic at the time of detection as they claim, there would be no benefit from early surgery (3).

I conclude that there is no difference in cancer dynamics that could be inferred from these incidence curves, and that early detection of breast cancer is still the best option in order to shorten time to surgery. However, a 30% increase of primary breast cancer after implementation of x-rays screening is worrying and suggests that x-ray screening may be carcinogenic, at least in women with hereditary DNA repair defect (4). A retrospective study comparing the percentage of DNA repair defects in metastatic breast cancer before and after initiation of widespread mammography would be useful.

1) http://www.ncbi.nlm.nih.gov/pubmed/26659430 2) http://seer.cancer.gov/ 3) http://www.ncbi.nlm.nih.gov/pubmed/15126770 4) http://www.ncbi.nlm.nih.gov/pubmed/22788243