Background: Clinically reliable predictive assays for normal tissue ra
diation sensitivity would help to avoid severe radiation induced morbi
dity and result in individualized dose prescriptions. Profound differe
nces of individual fibroblast and lymphocyte radiation sensitivity in
vitro have been documented in patients with certain genetic syndromes
but also in patients without known genetic disorders. The following re
view evaluates whether fibroblast or lymphocyte radiation sensitivity
measured in vitro correlates with the degree of acute and late radiati
on induced morbidity. Results: Acute radiation side effects and lympho
cyte sensitivity has been investigated in 2 studies. One of them repor
ted an insecure correlation, the other no correlation at all. Fibrobla
st radiation sensitivity and the extent of acute radiation induced sid
e effects on skin and mucosal sites has been compared in a total of 5
studies. None of these studies found a consistent significant correlat
ion Lymphocyte radiation sensitivity and late effects have been studie
d by 2 institutions. Late radiation induced skin and mucosal changes d
id not correlate with lymphocyte sensitivity in head and neck cancer p
atients, whereas in breast cancer patients a weak (R-2 = 0.06) correla
tion between the degree of late skin reactions and lymphocyte sensitiv
ity was observed. Late skin or mucosal radiation reactions and fibrobl
ast sensitivity were examined by 5 research groups. Data analysis reve
aled significant correlations or at least a trend towards a significan
t correlation in all studies. The quality of the reported correlations
expressed as R-2 ranged from 0.13 to 0.60, indicating a low predictiv
e value. Conclusions: Lymphocyte radiation sensitivity as measured by
currently available assays does not or only poorly correlate with acut
e and late effects of radiation in patients, precluding predictive tes
ts based on lymphocyte sensitivity. Fibroblast radiation sensitivity d
oes not correlate with acute but generally correlates with late radiat
ion morbidity. The quality of the later correlation is insufficient fo
r a reliable predictive lest with currently available methods to deter
mine cellular radiation sensitivity.