Myelodysplastic syndromes among atomic bomb survivors in Nagasaki: similarities to and differences from de novo and therapy-related cases Authors: Koji Ando, Yasushi Miyazaki Published: 2024 ABSTRACT AbstractEpidemiological studies for atomic bomb (A-bomb) survivors clearly demonstrated that A-bomb radiation increased the risk of hematological neoplasms, such as acute and chronic leukemia, and myelodysplastic syndromes (MDS) among survivors. Several studies on MDS among survivors investigated its characteristics, and it seems that MDS among survivors has different features from those seen in de novo MDS and therapy-related MDS. In this short review, we describe the differences of clinical features, chromosomal alterations and genome aberrations among them.Keywords: myelodysplastic syndromes (MDS), atomic bomb (A-bomb) radiation, karyotype, gene mutations Epidemiological studies carried out on atomic bomb (A-bomb) survivors demonstrated that radiation from an A-bomb significantly increased survivors’ risk of hematological neoplasms, such as acute and chronic leukemia and myelodysplastic syndromes (MDS) [1, 2]. We focus on MDS found in some survivors, and describe its clinical features, chromosomal alterations and genome aberrations. MYELODYSPLASTIC SYNDROMES AND ITS ETIOLOGY MDS, one of the myeloid neoplasms (MN), was not generally recognized before the 1980s until the French-American-British (FAB) group proposed its morphological diagnostic criteria and classification in 1982 [3]. MDS is characterized by: (i) cytopenia due to ineffective hematopoiesis, (ii) dysplasia of hematopoietic cells (HSC), and (iii) tendency to transform into acute myeloid leukemia (AML) [4]. The etiology of MDS is not well understood, but age-related clonal hematopoiesis (without hematological symptoms), called ‘clonal hematopoiesis of indeterminate potential (CHIP)’ [5] is assumed as an important stage prior to the development of clinically identified, usually symptomatic MDS. MDS is a disease of the elderly [4], and it is shown that genome alterations accumulated in HSCs are major causes of MDS, though no specific alterations are identified [6]. As mentioned, age-related genome alterations have a strong relationship with the development of CHIP, then MDS which is called de novo MDS. It is well known that the exposure to cytotoxic agents and radiation to treat malignancies increases the risk of MN which is called therapy-related MN (tMN) including therapy-related AML (tAML), MDS (tMDS) and myelodysplastic/myeloproliferative neoplasms (tMDS/MPN) [7]. For example, patients who have a history of chemo-radiotherapy for their primary malignancies (e.g. breast cancer, colon cancer, lymphoma) are at higher risk of developing tMN than those without the history. Chemotherapeutic agents are known to cause tMN by itself, however, the contribution of radiation alone has not been clearly shown, because most patients who received radiotherapy are also treated with chemotherapeutic agents during their management procedures. There are several reports that investigated the effects of medical radiation on the development of tMN analyzing patients who were treated radiation therapy (RTX) only (no chemotherapy). One study compared the risk of tMN among patients with well-differentiated thyroid carcinoma treated with surgery alone and with radioactive iodine [8]. They analyzed 148 215 patients, and found that radioactive iodine treatment was associated with an increased risk of developing AML and chronic myeloid leukemia (CML) than surgery alone by multivariate analysis. In this study, MDS was not found to be related to radiotherapy. Another study focused on the development MDS after external RTX for prostate cancer [9]. They investigated patients with prostate cancer who were treated by surgery (prostatectomy) alone and by external beam radiotherapy. Among 10 924 cases, they identified 31 cases who later developed MDS. However, radiotherapy was not selected as a risk factor of MDS by multivariate analysis. So far, the risk of tMDS after exposure to medical radiation alone has not been clearly shown via epidemiological studies. Other than chemo-radiotherapy, exposure to some chemicals such as benzene are also known to increase the risk of MN. Epidemiological data for workers in factories handling benzene showed its relationship including AML and MDS, however, this risk is negligible now under the current environmental regulation for benzene and related chemicals [10]. In general, 10–20% of newly diagnosed MDS (and also AML) patients are therapy-related, having a history of chemo-radiotherapy for their primary malignancies [11, 12]. Alkylating agents (AL-a) are well-known to be related with tMN, and also topoisomerase inhibitors (TI-i). Both AL-a and TI-i generate DNA damage, leading to cell death, but this sometimes results in the transformation of damaged cells. This is one of the reasons why the risk of MN increases after chemo-radiotherapy [13]. Other than this, the selection of already mutated HSC is another mechanism for the development of tMN [14]. Aged people have some HSCs which already have some genome aberrations (e.g. having age-related DNA mutations). If the mutated HSCs have a little resistance to chemo-radio therapy or a little growth advantage than normal HSC, chemo-radiotherapy and following transient suppression of normal hematopoiesis would become an opportunity for mutated HSCs to expand their clone size. This ‘bottleneck effects’ could be an initial step for tMN. The microenvironment of bone marrow consists of osteoblasts, adipocytes, fibroblasts, etc., and is essential for the maintenance, growth and differentiation of HSCs. Damaged microenvironment by chemo-radiotherapy are also thought to affect the development of tMNs [15, 16]. CHARACTERISTICS OF tMN (tMDS) Clinical features of tMDS are different from those of de novo MDS [7]. Usually, the prognosis of tMDS is worse than that of de novo cases, associated with frequent complex chromosomal abnormalities involving chromosome 5 and 7 alterations (monosomy or deletion of their long arms). Morphological dysplasia is evident, and clinical risk assessed using clinical scoring systems, such as International Prognostic Scoring System (IPSS) [17], or its revised version (IPSS-R) [18], is usually high. Recent genome analyses demonstrated high mutation rates of the TP53 and PPM1D genes, associated with complex karyotype [19]. Among tMDS, the prognosis of patients with TP53 mutations are generally poor, with low remission rates, and a short duration of remission [20]. Some investigators reported that even allogeneic stem cell transplantation cannot overcome its resistance. EPIDEMIOLOGICAL STUDY FOR MDS AMONG SURVIVORS Etiologically, it has been shown that A-bomb radiation increased the risk of AML [1]. Considering the impact of chemo-radiotherapy on the development of tMN including tAML and tMDS, we were interested in the effects of A-bomb radiation on the risk of MDS among survivors. Radiation Effect Research Foundation (RERF), formally Atomic Bomb Casualty Commission (ABCC), started the Life Span Study (LSS) in 1950 [21]. This is a cohort study of A-bomb survivors to investigate the late effects of A-bomb radiation on human health including the incidence of malignancies. This study is ongoing, and has been providing important data regarding the risk of radiation on the development of neoplasms. The risk of radiation on human health estimated through LSS have been widely used to make the standards for radiation protection. The risk of leukemias among A-bomb survivors was shown to be elevated using the data of LSS [1], however, MDS was not included in the target diseases of LSS when started in 1950. MDS was not well recognized until the late 1980’s even among hematologists, as mentioned above. In the LSS, data has been collected using International Classification of Disease for Oncology (ICD-O), in which MDS became coded as ‘malignant’ in the 3rd edition released in 2011. These were probably the reasons why MDS data has not been collected until recently in LSS, and we needed to perform another epidemiological study to see the relationship between A-bomb radiation and the development of MDS [2]. Nagasaki University Hospital and affiliated institutions in Nagasaki city have organized ‘Nagasaki open city cohort’ (NOCC) for MDS since 1985 which covers approximately 80–90% of newly diagnosed MSD cases in Nagasaki city [2]. By using database for survivors collected in Atomic Bomb Disease Institute (ABDI, established in Nagasaki University), and NOCC data, we could identify MDS cases among survivors diagnosed in Nagasaki. In the ABDI database for about 120 000 cases in total, there were 64 058 survivors in Nagasaki city who were alive in 1985, and we identified 151 MDS patients in NOCC out of 64 058 survivors diagnosed between 1985 and 2004. Although estimated exposure dose was not available for cases in NOCC data, we had information regarding the exposure distance from the hypocenter for survivors. According to the dosimetry system published in 2002 (DS02), which estimated the radiation dose released by the A-bombs, the doses of A-bomb radiation at 1.5 km and 3 km from the hypocenter were roughly 1 Gy, and 5 mGy, respectively, in Nagasaki (without shielding). Among MDS survivors in NOCC data, after adjusting sex and age at exposure, MDS incidence rate showed statistically significant inverse relationship with the exposure distance. Comparing the LSS data and NOCC data, we also identified MDS cases in LSS Nagasaki database. There were 44 MDS patients in the LSS-Nagasaki cohort who were alive at 1985 (among 22 245 survivors), and diagnosed as MDS by 2004. LSS data contained estimated bone marrow dose using DS02 for each cohort member, and also the exposure distance from the hypocenter. Excess relative risk (ERR) of MDS showed a linear relationship with bone marrow dose (ERR 4.3 at 1 Gy), which was larger than those for acute lymphoblastic leukemia and AML, and also a significant inverse relationship with the exposure distance. These results clearly demonstrated that A-bomb radiation significantly increased the risk of MDS among survivors in Nagasaki. It was suggested that the dose of radiation and distance from the hypocenter had stronger effects on the development of higher-risk MDS than lower-risk MDS. CLINICAL FEATURES OF MDS AMONG SURVIVORS INCLUDING THE CYTOGENETIC CHARACTERISTICS To investigate whether MDS among survivors have different clinical features from de novo MDS, we collected MDS cases in Nagasaki, and confirmed the history of exposure to A-bomb radiation in each case [22]. There were 269 patients unexposed (de novo MDS), and 133 patients among survivors. In this comparison, we excluded cases who had a history of chemo-radiotherapy before the diagnosis of MDS to exclude possible tMDS. There were no clear differences in age, male/female ratio, FAB subtypes, levels of cytopenia (Hb, neutrophil count, platelet count), and bone marrow blast percentage between de novo MDS cases and MDS among survivors. As mentioned above, clinical risk prediction for MDS is performed using the IPSS scoring system, categorizing MDS into four groups with different prognosis using cytogenetic risk group, levels of cytopenia and bone marrow blast percentage. The IPSS risk category of the exposed and unexposed groups was not different (Table 1). The percentage of cases with abnormal karyotype was similar between two groups, and also the distribution in revised IPSS (IPSS-R) cytogenetic risk categories. Next, we further divided MDS survivors by the distance from the hypocenter into three groups, because radiation dose represented by the distance could have an impact on the clinical and cytogenetic features of MDS among survivors. As in Table 2, there were 29 patients exposed within 1.5 km from the hypocenter (Group I), 35 between 1.5 and 3 km (Group II), and 69 exposed at more than 3 km (Group III). Among three groups, platelet count was high in Group I, and the IPSS high risk patients were more frequent in Group I (20.7%, 8.6% and 5.8% in Group I, II and III, respectively). The most striking difference among three groups existed in karyotype; abnormal karyotype was observed at 72.4% of patients in Group I, in which Very poor IPSS-R cytogenetic category shared 27.6% (8.6% and 7.2%, in Group II and III, respectively. Table 2). This difference strongly suggested poor prognosis in Group I, however, there was no statistical difference in overall survival (OS) for MDS between survivors and unexposed cases (P = 0.39), or leukemia-free survival (P = 0.72). Considering the largest impact of karyotype on survival for MDS, and significantly more frequent high risk IPSS-R karyotypes in Group I, we thought that there might be a difference in cytogenetic features of MDS among survivors. Table 1Comparison of clinical and cytogenetic features of MDS among survivors and de novo MDS [22]. A-bomb survivors totalUnexposed n%n%Total patients133269Sex Male7153.416762.1 Female6246.610237.9Age (Years) <60107.55420.1 ≧6012392.521579.9 Median7272 Range42–9441–92Diagnosis period 1985–19941914.32910.8 1995–20047254.111442.4 2004–20134231.612646.8FAB classification RA9168.419271.4 RARS43.0103.7 RAEB2821.15319.7 RAEB-t43.093.3 CMML64.551.9Bone marrow blast (%) <59672.220576.2 5–101410.53412.6 11–201712.8238.6 21–3043.072.6CytopemiasHb, g/dL Median8.89 Range3.2–182.5–15.3ANC (×103/μL) Median1.41.5 Range0.1–31.70.1–24PLT (×103/μL) Median9082 Range1–8580–544Observation time (months) Median4025 Range0–2540–330IPSS Low risk3123.35420.1 Intermediate 16851.114553.9 Intermediate 22115.85721.2 High risk139.8134.8Karyotype Normal7153.415256.5 Abnormal6246.611743.5IPSS-R cytogenetic risk Very good64.5124.5 Good8060.217063.2 Intermediate2216.53613.4 Poor96.8259.3 Very poor1612.0269.7 Comparison of clinical and cytogenetic features of MDS among survivors and de novo MDS [22]. Table 2Clinical and cytogenetic features in groups by exposure distance from the hypocenter in MDS among survivors [22]. <1.5 km (Group I)1.5–2.99 km (Group II)3 km- (Group III) n%n%n%Total patients293569Sex Male1655.22262.93347.8 Female1344.81337.13652.2Age (Years) <60310.3411.434.3 ≧602689.73188.66695.7 Median747271 Range54–8349–9042–94Diagnosis period 1985–1994413.8617.1913.0 1995–20041862.11954.33550.7 2004–2013724.11028.62536.2FAB classification RA1965.52571.44768.1 RARS00.012.934.3 RAEB724.1514.31623.2 RAEB-t26.925.700.0 CMML13.425.734.3Bone marrow blast (%) <51965.52777.15072.5 5–10413.825.7811.6 11–20413.825.71115.9 21–3026.925.700.0CytopeniasHb, g/dL Median8.28.89.1 Range5.5–12.65.6–14.63.2–18ANC (×103/μL) Median1.3521.3 Range0.1–31.70.2–7.50.1–6.3PLT (×103/μL) Median1068489 Range29–4344–8581–440Observation time (months) Median463437 Range3–2070–2540–252IPSS Low risk413.81028.61724.6 Intermediate 11551.71645.73753.6 Intermediate 2413.8617.11115.9 High risk620.738.645.8Karyotype normal827.61851.44565.2 abnormal2172.41748.62434.8IPSS-R cytogenetic risk Very good00.025.745.8 Good931.02160.05072.5 Intermediate827.6822.968.7 Poor413.812.945.8 Very poor827.638.657.2 Clinical and cytogenetic features in groups by exposure distance from the hypocenter in MDS among survivors [22]. To address this issue, we performed a detailed cytogenetic analysis comparing Group I and unexposed cases (Table 3) [22]. In Group I, the most frequent chromosomal aberration was deletions of chromosomes (observed in 13 out of 21 cases, 61.9%), followed by structural gain (11 cases, 52.4%), structurally neutral changes, such as translocations and inversions (10 cases, 47.6%), monosomy (seven cases, 33.3%) and trisomy (five cases, 23.8%). Among these aberrations, random balanced translocations and inversions were significantly increased in Group I (12 out of 123 cases, 9.7%, P < 0.001). The frequency of others including deletions were similar between two groups. Table 3Comparison of the frequency of the type of chromosomal aberration [22]. Group I (n = 21)Unexposed (n = 123) P value n%n% Monosomy−1466.77661.8+733.34738.20.809Trisomy−1676.29577.2+523.82822.80.999Deletion−838.15746.3+1361.96653.70.636Structural gain−1047.68065.0+1152.44335.00.148Structural neutral change−1152.411190.2+1047.6129.8<0.001 Comparison of the frequency of the type of chromosomal aberration [22]. Ionizing radiation damages DNA, generating base damage, single-strand breaks and double strand breaks (DSB) of DNA. Among them, DSB characterizes the damage caused by ionizing radiation, and it is assumed that deletions, translocations and inversions would be the major chromosomal consequences of exposure related-mismatch repair of DSB. It has been demonstrated that stable chromosomal translocations were observed in peripheral blood leukocytes of heavily exposed survivors without hematological changes [23], and it is possible that some of these translocations present in MDS of survivors be attributable to A-bomb radiation. These changes might have different prognostic meaning for MDS patients than chromosomal abnormalities usually observed in de novo cases. In fact, structural neutral changes including translocations and inversions are rare in de novo MDS. We next checked whether karyotype alterations were accumulated in specific chromosomes in Group I cases. As radiation penetrates into cells randomly, ionizing radiation would damage chromosome 1, the largest chromosome, more frequently than others. Indeed, deletions in chromosome 1 were more often present in Group I than in the unexposed group (P = 0.01), and the incidence of deletions in chromosomes 5 and 20 (both are frequent in de novo MDS cases) was equal. Monosomy of chromosomes 5, and 7 are frequently observed in MDS, especially in tMDS, but their frequency in Group I was not significantly different from those in the unexposed group. Structural gain in chromosome 11 was significantly increased in Group I (P = 0.002). Combining structural gains, deletions and structural neutral changes, chromosome 11 was affected in seven out of 21 cases (33.3%) in Group I. In the unexposed group, chromosome 11 was damaged in six out of 123 cases (4.9%), and there was a statistically significant difference between them (P = 0.001). Chromosomes 3 and 8 were also affected significantly more often in Group I (six and three, out of 21 cases, respectively) than in the unexposed group (eight and one, out of 123 cases, respectively, P = 0.007 and 0.01, respectively). Although these cytogenetic features existed, IPSS-R cytogenetic risk categories could stratify MDS patients among survivors, showing clinical impact on survival. However, it was suggested that Very poor cytogenetic category, defined only by the number of chromosomal aberrations (>3) [18, 24], had different impact on OS for survivors’ MDS, which had more structural neutral changes. Through this study, we understand that MDS among survivors had different clinical and cytogenetic features than tMDS, such as survival (OS for tMDS is usually dismal), aberrations in chromosome 5 and 7 (frequent in tMDS than de novo MDS), and types of aberrations (translocation is not common in tMDS). Karyotypes of MDS among survivors were not simply characterized by the presence of (stable) structural neutral changes probably generated by A-bomb radiation, because it does not fully explain why the incidence of MDS increased among survivors. In this regard, frequently affected chromosomes (numbers 3, 8 and 11) might have some roles for the pathogenesis of MDS after exposure to A-bomb radiation. In terms of the treatment response for MDS among survivors, unfortunately, we did not have many patients who were managed with current standard treatment, such as lenalidmide, erythropoiesis stimulating agents, or azacitidie. We could not evaluate the efficacy of these drugs on MDS among survivors. Allogeneic HSC transplantation has not been performed for them at least in Nagasaki, mainly because of the high age of this group. Cytogenetic analyses demonstrated specific feature of MDS among survivors, such as increased incidence of structural neutral changes, and frequent aberrations in chromosome 3, 8 and 11. The number of chromosomal aberrations seemed to have different clinical impact than de novo MDS [22]. There is a report regarding AML and MDS after RTX alone (RTX-AML/MDS) for solid cancers, such as prostate and breast cancer [25]. The authors of this report concluded that RTX-AML/MDS patients had different clinical characteristics from those after chemotherapy with or without RTX, pointing out the cytogenetic features and clinical behavior of RTX-AML/MDS were similar to those of de novo AML/MDS. It was assumed that RTX-AML/MDS may not represent a direct consequence of radiation. In their data, cytogenetic risk distribution was not different between RTX-AML and de novo AML, and also RTX-MDS and de novo cases, but AML/MDS after chemotherapy had different cytogenetic profiles. In this report, the increased incidence of RTX-AML or -MDS was not clearly shown, and we do not know whether ‘RTX-AML/MDS’ in this report was etiologically “radiation-induced” or not. These observations may not be simply applied to MDS among survivors, for example, median duration between RTX and the onset of MDS was 60 months in their study, which is shorter than that for MDS among survivors in our series (diagnosed at least 40 years after the exposure). As mentioned in this review, the incidences of AML and MDS among survivors were shown to have significant relationship with A-bomb radiation, and they had different cytogenetic features from de novo cases. In terms of survival, at least for MDS, there seemed no significant difference between A-bomb related and de novo patients, but it did not prove that RTX-MDS in this report and A-bomb related MDS had similar pathogenesis. GENOME ANALYSIS FOR MDS AMONG SURVIVORS To further address the pathogenesis of MDS among survivors, we analyzed bone marrow and peripheral blood samples using next generation sequencing technologies [26]. We obtained samples from 32 patients diagnosed as having MDS, and three patients as idiopathic cytopenia of undetermined significance (ICUS) among survivors. These 35 patients were divided into two groups, proximally exposed group (PE) and distantly exposed group (DE). Patients in the PE group were exposed to A-bomb radiation within 2.7 km from the hypocenter. Estimated dose at 2.7 km was about 5 mGy, and each estimated dose for those in the DE group was assumed to be less than 5 mGy. Whole genome sequence (WGS), whole exome sequence (WES) and target sequence (TS) were applied depending on the availability of genome DNA sample for each case. Some samples were also analyzed using SNP array to detect DNA copy number alterations. There was no major difference in clinical features (median age [73- and 74-years-old in PE and DE group, respectively], WHO subtype of MDS, median age at exposure [both 12-years-old]) between PE and DE groups. Median exposure distance from the hypocenter was 1.1 km in PE, and 3.4 km in DE group. By using three methods (WGS, WES and TS), somatic and oncogenic mutations were identified in 16 out of 18 patients (89%) in the PE group, and 12 out of 17 patients (71%) in the DE group (Fig. 1). Among these mutations, in the DE group, TET2 was most frequently affected (five out of 17 patients, 29%), followed by SF3B1 (three out of 17, 18%) and STAG2 (18%). Mutations in TP53 were found in two cases in the DE groups (12%). These genes are often mutated in de novo MDS, and their frequencies were not largely altered from generally reported numbers [24, 25]. On the other hand, in PE group, none had TET2 or STAG2 mutations, and the most frequently mutated gene was SF3B1 (four out of 18 patients, 22%), whose percentage was similar to that in DE group. The frequency of TET2 mutations was significantly decreased in PE than DE group (P = 0.019), but STAG2 was not (P = 0.104). Mutations in TP53 were identified at 11% in PE group, which was very similar to that in DE group (12%). There was also no significant difference in the frequency of RUNX1 mutations between the two groups (11% and 6% in PE and DE, respectively; P = 1.00). There is a report studying on mutations of RUNX1 in MDS among survivors (including both proximally and distally exposed cases) in Hiroshima [27]. The investigators in Hiroshima found an increased alteration rate of 46% (six out of 13 cases), and a missense/frameshift mutation rate of 31% (four out of 13 cases), which were apparently higher than our results. In the present study, three out of 35 cases (8.6%) had RUNX1 mutations, which was as frequent as reported for de novo MDS (approximately 10%). There is no clear explanation for this difference in the frequency of RUNX1 mutations, but the small number of cases examined in each study (13 cases in the Hiroshima study, and 35 cases in this study) might have played a role. It is also possible that the differences in MDS subtypes in two studies influenced the results. Mutations in the RUNX1 gene are generally enriched in higher-risk MDS. [27] In the report from Hiroshima, the diagnosis of one patient was refractory anemia with excess blasts (RAEB), eight with RAEB in transformation (RAEB-t), and one with AML, resulting in AML and MDS with increased blasts at 76.9% (10 out of 13), of which Nagasaki study included 12 out of 35 cases (34.3%). Somatic mutations in MDS among A-bomb survivors. Somatic mutations were detected using whole genome sequencing, whole exome sequencing, and target sequencing [26]. Mutated genes were categorized on the basis of their assumed roles in functional pathways (Fig. 2). Major mutations in the DE group were genes for DNA methylation, RNA splicing and transcription. However, in PE group, mutations in the DNA methylation pathways (including TET2) were significantly less (one out of 18 patients, 5.6%) than in DE group (seven out of 17, 41%; P = 0.018). We did not find a significant difference in mutations of genes coding RNA splicing factors. Aberrations in genes for transcription factors and the chromatin modification pathway were more frequent in PE than in DE without statistical significance (transcription factors, 39% and 24%, respectively, P = 0.47; chromatin modification pathway, 33% and 12%, respectively, P = 0.23). Frequencies of mutated genes categorized by assumed functional pathway. Mutated genes were categorized by putative functional pathways. There was a statistically significant difference in DNA methylation pathway between the PE and DE groups shown with ‘*’ (P = 0.018). PE, proximally exposed; DE, distally exposed [26]. Through this study, we found that in MDS among survivors, the gene mutation profiles were different between proximally exposed and distantly exposed patients. It was suggested that dose of A-bomb radiation had effects on the development of MDS by inducing a different mutation profile than de novo cases. QUESTIONS REMAIN MDS derives from HSCs, that generate hematopoietic cells through life. Our studies on MDS among survivors, and also other genetic and cytogenetic studies support the idea that A-bomb radiation affected HSCs, providing genome aberrations. Recent studies regarding the mutational signatures of radiation-induced cancers have revealed small deletions and structural neutral change of chromosomes, such as inversions, are candidate radiation-signatures of cancers [28, 29]. De novo MDS is thought to arise from preceding age-related clonal hematopoiesis, or CHIP [5]. It will be necessary to study CHIP in survivors, and this will add further insights into the pathogenesis of radiation-induced MDS. COI STATEMENT There is no conflict of interest to disclose. Contributor Information Koji Ando, Department of Hematology, Atomic Bomb Disease and Hibakusha Medicine Unit, Atomic Bomb Disease Institute, Nagasaki University, Nagasaki 852-8523, Japan. Yasushi Miyazaki, Department of Hematology, Atomic Bomb Disease and Hibakusha Medicine Unit, Atomic Bomb Disease Institute, Nagasaki University, Nagasaki 852-8523, Japan. REFERENCES