As far as I know, it is reliable enough to still be the "Gold standard" in radiation biodosimetry. The dicentrics are radiation-specific, and their biological "half-life" is known so one can back-calculate doses. The method is fairly low-tech (which can be good as one does not have to rely on expensive machines) but nowadays automatic systems are available that scans the slides for metaphases (as opposed to a person having to scan the slides for metaphases). On the downside is the lag period after which one can obtain results (the blood must be cultured a few days), and the already mentioned time consuming scanning of the slides to find metaphases.

To some extent, a partial-body exposure can also be detected by studying the distribution of aberrations in the cell population.

I hope some of the real experts can fill in the blanks here, I can't claim to have all the latest updates  :)

In such situation L propose use LTN-method. Lympphocytes with Tailed Nuclei (LTN) are formed from dicentrics after bribged mitosises. You may see LTN appearence in peripheral blood lymphocytes after irradiation. More information concerning LTNs in bioindication of radiational expose see in my profile (last paper)

I'd like to continue Stephen Barnard's reply. Dicentrics assay is the only method of biodosimetry, for which ISO exists (actually, two ISO). The "limited number of dicentrics formation" has a double meaning for me. Meaning Nr One: there is a limited number of chromosomes in the lymphocyte, so there is a certain physical limit for the number of dicentrics, which can be formed (no more than 21-22). This might happens at extremely high radiation doses and can be observed as a saturation of the dose response curve starting from 20 Gy and above. The most extensive study was published by M. Sasaki, 2003, IJRB.

Meaning Nr Two: The yield of dicentrics you observed in vivo was too low. There are two ways to dealing with this problem. One is what Stephen suggested - you have to score more cells. The question of dose detection limits and dose measurement accuracy was comprehensively examined and presented in papers by Lloyd et al (2006) The Minimum  Detectable Dose by Biodosimetry in a Radiation Overexposure and Szłuin´ska et al (2007) PRESENTING STATISTICAL UNCERTAINTY ON CYTOGENETIC DOSE ESTIMATES. In my opinion, the detection limit for dicentric assay is about 20-30 mGy gamma-rays (a doubling dose) and the practically achievable measurement threshold is about 50-70 mGy. I think, this is the best we can get with classic statistical methodology.

Alternatively, we can apply Bayesian analysis or Bayesian-like technique for data treatment, and get a dose estimate in a probabilistic manner. This approach helps even if the dicentric yield in exposed person in close to zero. We applied it over a decade ago to Chernobyl clean-up workers, and generated rather informative distributions of radiation doses in  this cohort. At the moment the world top expert in Bayesian analysis applications to chromosomal biodosimetry is Stephen's boss, Dr. Elizabeth Ainsbury of the CRCE, PHE, UK. 

More comments are available from me by the request.

Thanks Volodymyr for your answer. I meant the meaning No. 2 described by You.

I agree that the dicentric chromosome assay is reliable. However, I do not think that dicentrics are always specific to radiation and a half life of dicentrics is necessarily known. It may also be important to discuss the purpose of biodosimetry (e.g., triage or retrospective biodosimetry), depending on which the FISH translocation assay may be more reliable, or other approach such as the micronucleus assay may be more useful, than dicentric chromosome assay.

https://www.ncbi.nlm.nih.gov/pubmed/27431686

https://www.ncbi.nlm.nih.gov/pubmed/25922373

https://www.ncbi.nlm.nih.gov/pubmed/25520376

https://www.ncbi.nlm.nih.gov/pubmed/11382333

CA test is highly reliable at doses greater than 250 mGy or so. At lower doses say >100 mGy or so, the reliability is poor. Please note that 1 dicentric may sometimes be present in 500 scored cells as a background (due to smoking, pollution, food habits, body mutations etc. etc) and roughly ~(65-85) mGy or so, is required to get 1 dicentric. Hence at doses of the order of 100 mGy, one commonly gets 1 dicentric or 2 if 500 cells are scored. The above explanation is for low LET radiations associated with Cs or Co sources or so.  AS mentioned by Carl as well as N. Hamada, it is a gold  and reliable standard in Bio dosiemtry. Better to perform CA test as early as possible if suspicions arise or TLD/OSLD reports high doses >100 mSv. For

The Special Issue on the European RENEB project guest edited by Drs Ulrike Kulka and Andrzej Wojcik has just become available online as the January 2017 issue of International Journal of Radiation Biology. All papers in this special issue are freely downloadable.

http://www.tandfonline.com/toc/irab20/93/1

Please refer the attached file where you can see the applications of Micronucleus Assay, Chromosomal Aberrations and Comet assay. 

Article Antigenotoxic Potential of Rutin and Quercetin in Swiss Mice...