EPA's response refers to the various advisory body reports stating "Again and again, these bodies have endorsed LNT as a reasonable approach to regulating exposures to low dose radiation.". It then refers to BEIR VII report.
These advisory bodies have ignored available evidence and have utilized faulty evidence to claim support for the LNT model. See more details in "Deficiencies of advisory bodies".
While referring to evidence since the BEIR VII report, EPA's response stated: "In this connection, we would note, inter alia, results of epidemiological studies on: nuclear workers in the United States, France and the United Kingdom ; residents along the Techa River in Russia who were exposed to radionuclides from the Mayak Plutonium Production Plant [6,7]; and children who had received CT scans . These studies have shown increased risks of leukemia and other cancers at doses and dose rates below those which LNT skeptics have maintained are harmless - or even beneficial."
 - INWORKS study has invalid study design. Occupational dose does not represent total radiation dose to the workers due to the increased medical radiation dose in later years (Doss, 2015).
[6,7] - Techa River cohort studies have the issue of increased medical attention to the population that had higher doses (Jargin, 2014)
 - (Pearce et al., 2012) has poor study design. Reverse causation is likely the reason for the observed results (Boice, 2015)
Here are brief comments to some of the items in NCI's response:
Atomic bomb survivor data - NCI referred to the atomic bomb survivor data's characteristics and claimed they supported the LNT model, but did not respond to the analysis by (Doss, 2013) which showed the data are more consistent with radiation hormesis.
Radiation worker studies - NCI discounted the observed reduction of cancers in these radiation worker studies by citing healthy worker effect (HWE) but did not refer to the Nuclear Shipyard Worker Study which used non-radiation workers as controls and so avoided the HWE issue. Also, it referred to INWORKS studies which have a basic study design flaw (Doss, 2015).
TB Fluoroscopy cohort - NCI referred to the later study by Howe which did not show the hormetic reduction of breast cancers in the ~12 cGy region. However, in this study, much larger dose bins were used and so effectively masked any cancer reduction. The Massachusetts study also used large bin size and so had a similar issue. These studies do not show increased risk of cancer at low doses.
2006 Taiwan study: NCI stated "The petitioners fail to mention a later and more reliable internal analysis of this cohort" and referred to the 2008 study. The observed reduction of 2006 study in comparison to an age-matched group continues in these updated data (Doss, 2013). The 2008 study shows significant increase in two cancer types only because the authors lowered the confidence interval used, from 95% used in 2006 publication to 90%. If they had continued to use 95% CI in 2008, the increase would not be significant.
NCI also quoted the already discredited (Pearce et al, 2012) study which has poor study design. Reverse causation is likely the reason for the observed results (Boice, 2015)