Feasibility of thoracic CT in assessing anemia for aplastic anemia patients undergoing allogeneic hematopoietic stem cell transplantation

Abstract

BACKGROUND:

Anemia is an important clinical symptom for aplastic anemia (AA) patients who are suffered with peripheral pancytopenia.

OBJECTIVE:

To evaluate the accuracy of diagnosing anemia with non-invasive chest computed tomography (CT) for AA patients.

METHODS:

The CT attenuation of left ventricular (LV) cavity and interventricular septum (IVS) on unenhanced thoracic CT images of AA patients are retrospectively analyzed, including 84 AA patients in pre-transplant and 1-month (n = 82), 2-month (n = 72), 3-month (n = 75), 6-month (n = 74) and 12-month (n = 70) followed patients in post-transplant. The difference (IVS-LV) and ratio (LV/IVS) of the CT attenuation between LV cavity and interventricular septum are calculated. Serum hemoglobin is estimated within 24 hours of CT imaging. The CT attenuations of IVS-LV and LV/IVS are correlated with hemoglobin, and their variation tendency is analyzed during the treatment of a-HSCT. A receiver operating characteristic (ROC) curve analysis is then performed for the diagnosis of anemia.

RESULTS:

The CT attenuations of IVS-LV and LV/IVS well correlate with hemoglobin (r = –0.618 and 0.628, respectively, P < 0.001). The variation tendency of IVS-LV and LV/IVS is similar to that of hemoglobin with opposite directions during one-year follow-up of a-HSCT. When a threshold of CT attenuation of IVS-LV and LV/IVS is set at 11.5HU and 0.77, respectively, both the sensitivity and specificity in diagnosing anemia are good (74.7% and 73.8% in CT attenuation of IVS-LV; 77.4% and 70.4% in LV/LVS, respectively).

CONCLUSIONS:

Both CT attenuation of LV/IVS and IVS-LV had similar accuracy in diagnosing anemia for AA patients. The non-invasive chest CT can offer a new possibility to complementarily evaluate anemia for AA patients in the diagnostic radiology reports.

Keywords

Anemia computed tomography allogeneic hematopoietic stem cell transplantation hemoglobin aplastic anemia

1 Introduction

Aplastic anemia (AA) is a rare and life-threatening hypocellular bone marrow that results in hematopoietic failure and pancytopenia [1, 2]. The disease occurs more frequently in Asia, with 2- to 3-fold higher incidence rates than in the West [3]. Allogeneic hematopoietic stem cell transplantation (a-HSCT) is the first choice of treatment in patients under 40 years old with aplastic anemia who possess HLA-matched sibling donors [4]. One of the most important clinical symptoms in AA patients is anemia, which can reduce tissue oxygen delivery and cause a compensatory cardiovascular response, including paleness, fatigue and syncope [5]. As we known, blood test is the current standard of screening for a load of blood components, which anemia is evaluated by the level of serum hemoglobin.

Many previous studies had shown that anemia could be indicated on unenhanced CT of the thorax [6 –11]. The measurement of CT attenuation difference (IVS-LV) between LV cavity and interventricular septum had been demonstrated to be superior to the LV cavity, which was considered as the effectively quantitative measurement of diagnosing anemia by CT scans. However, the causes of anemia in those studies were unknown or different, such as lymphoma, chronic nephritis, malignant tumors, et al. In addition, neither they had large sample or comparative study after treatment, nor do the reports compared the accuracy of ratio (LV/IVS) of CT attenuation between LV cavity and interventricular septum with IVS-LV, which was supposed to make more stable effect in diagnosis of anemia, theoretically.

Chest CT examination is frequently used in AA patients to exclude pulmonary infections and non-infectious complications, particularly during the treatment of a-HSCT. Besides the assessment of lung, it can provide an opportunity to quantify the CT attenuation of LV cavity and IVS in the same field of view. Some radiologists may comment on the density of the LV cavity and infer an anemic state for AA patients, as well as anemia caused by other diseases. However, AA is a special hematological system disease with peripheral pancytopenia and bone marrow hypoplasia, including decreased hemoglobin, platelet and neutrophil counts. AA patients usually suffer with moderate to severe anemia in the pre-transplant and the first several months after transplantation. It is different from other anemic patients who only have mildly to moderately decreased serum hemoglobin. Furthermore, the serum hemoglobin of ‘healthy’ HSCT recipients is usually lower than the normal laboratory range and can be defined as ‘anemic’. Whether the CT threshold in diagnosing anemia for AA patients is consistent with those anemic patients caused by other diseases is still unknown.

Therefore, the purpose of this study was to evaluate the accuracy of CT attenuation of LV/IVS and IVS-LV in diagnosing anemia for AA patients. The variation tendency of hemoglobin, IVS-LV and LV/IVS was compared during the treatment of transplantation in one year follow up. Furthermore, the CT thresholds of IVS-LV and LV/IVS in diagnosing anemia for AA patients were attempted to obtain. It was not meant to assessing anemia compete against serum hemoglobin, but to offer a new possibility of getting valuable add-on information in a one-stop shop fashion in the diagnostic radiology reports, especially when the blood tests for AA patients are lacking or patients are not suitable for repeated blood tests due to the decreased platelet counts.

2 Methods

2.1 Study population

We retrospectively identified 94 consecutive AA patients from January 2017 to December 2018. With approval of the ethics committee of our hospital, the patients were informed that their medical information would be anonymized and used for research, although written consent was waived. The inclusion criteria were as follows: (1) The diagnostic criteria of AA must be at least two of the following [12]: hemoglobin concentration <100 g/L, platelet counts <50×10⁹/L, neutrophil counts <15×10⁹ /L; (2) the serum hemoglobin level including the pre-transplant time point (within 2 weeks before transplantation) and at least one of the post-transplant time points (1, 2, 3, 6 and 12 months after transplantation) were available; and (3) patients had unenhanced chest CT within 24 hours of hemoglobin examination.

The exclusion criteria were as follows: (1) didn’t had the treatment of a-HSCT; (2)had more than one times transplantation; (3) had no clinical or imaging data at the pre-transplant time point; (4) had no myeloid engraftment or died by one month; (5) had diseases that affect cardiac density, such as vasculitis lesions, glycogen accumulation, or hepatic haemochromatosis; and (6) images had motion or metal (calcification) streak artefacts. Finally, 84 AA patients in pre-transplant and those patients followed at 1-month (n = 82), 2-month (n = 72), 3-month (n = 75), 6-month (n = 74) and 12-month (n = 70) in post-transplant were all enrolled in the study (Detailed patients flow were depicted in Fig. 1). According to those AA patients’ serum hemoglobin level at different time points, 217 patients were anemia group, and the remaining 240 patients were non-anemia group. All patients were hospitalized in a laminar airflow room, underwent the conditioning regimen and were given prophylactic antibiotics before transplantation.

Fig. 1

The inclusion and exclusion chart at different time points for AA patients during the treatment of a-HSCT. Note: The green block diagrams represented the exclude patients and the red represented the included patients. AA: aplastic anemia; a-HSCT: allogeneic hematopoietic stem cell transplantation.

2.2 CT scan analysis

All non-contrast thoracic CT examinations were performed on 16-detector row CT scanner (Aquilion 16; Toshiba Medical Systems, Tokyo, Japan) from the thoracic inlet to diaphragm. The scan parameters for this protocol were a tube voltage of 120 kV, an automatic tube current of 10%,400 mAs, a rotation time of 0.5 s, and a transverse slice thickness of 5 mm. Unenhanced chest CT images were evaluated on a preset standard mediastinum window with the following Hounsfield Unit (HU) including window width of 400 HU and window level of 60 HU.

2.3 Image analysis

Two experienced radiologists who have experience of 10 and 15 years in cardiovascular diagnosis, respectively, and are blind to the subjects’ characteristics independently read and measured the CT value (HU) of patient’s LV cavity and interventricular septum on a Picture Archiving and Communication System (PACS). The LV cavity measurement levels were chose avoiding artefacts from the spine and the best to display interventricular septum. Regions of interest (ROIs) were placed on LV cavity with an area of approximate 100 mm², avoiding the inclusion of ventricular papillary muscles. The irregularly long strips base on the interventricular septum were measured as large as possible on the same section of the measuring of LV cavity (Fig. 2). The CT value of LV cavity and the interventricular septum measured by two radiologists were averaged and the value of IVS-LV and LV/IVS were calculated, respectively.

Fig. 2

(A) a 36-year-old male AA patient with anemia at pre-transplant time point (Hb level 59 g/L). A round ROI (area of approximately 110 mm²) was made to measure the CT value of LV cavity, avoiding the inclusion of ventricular papillary muscles. A long strip ROI (area of approximately 85 mm²) along the margin of the interventricular septum was drawn as large as possible, on the same section of the measuring of LV cavity. The CT value of LV cavity and interventricular septum was 19 HU, 33 HU; the value of IVS-LV and LV/IVS was 14HU and 0.57, respectively. (B) At the post-transplant time point of 6 months, the serum hemoglobin increased to normal (Hb level 131 g/L). The CT value of LV cavity and interventricular septum was 47 HU and 49 HU; the value of IVS-LV and LV/IVS was 2 HU and 0.95, respectively. LV: left ventricular; IVS-LV: difference between interventricular septum and left ventricular cavity; LV/IVS: ratio between left ventricular cavity and interventricular septum. Hb: hemoglobin.

2.4 Serum hemoglobin

The serum hemoglobin level was obtained by routine blood tests in clinical hematology laboratory at each time point: the pre-transplant time point (within 2 weeks before transplantation) and 1, 2, 3, 6 and 12 months after transplantation. According to the WHO and Chinese criteria of diagnostic anemia [11], serum hemoglobin(g/L) less than 60 g/L defined as severe anemia, (60–90) g/L defined as moderate anemia, (90–120) g/L in male or (90–110)g/L in female defined as mild anemia. Since AA patients usually are moderate to severe anemia before the treatment of transplantation and in the first several months after transplantation. The serum hemoglobin of ‘healthy’ HSCT recipients is usually lower than the normal laboratory range and can be defined as ‘anemic [13]. In addition, some transplanted matches of donors and recipients were different genders during the process of a-HSCT. Therefore, the critical level of anemia in our study was defined as hemoglobin level less than 90 g/L in both male and female.

2.5 Statistical analysis

The clinical characteristics (age, sex, serum hemoglobin) and measurement data (CT value of LV cavity, interventricular septum, IVS-LV, LV/IVS) were reported as mean±standard deviation. Inter-observer agreement of the two parameter measurements was assessed by reporting the intraclass correlation coefficient (ICC) and 95% confidence interval (CI). One-way analysis of variance (ANOVA) was used to compare differences among the continuous variables in different time points. Scatter plots and simple linear regression analyses were performed to identify the relationships of the observations (IVS-LV, LV/IVS) with peripheral blood hemoglobin level. In addition, the point plots were made to show the variation tendency of IVS-LV, LV/IVS and hemoglobin in one-year follow-up of a-HSCT. Finally, a receiver operating characteristic (ROC) curve analysis was used to evaluate the threshold CT value in diagnosing anemia. True-positive (TP) cases were defined as anemic patients correctly identified. True-negative (TN) cases were non-anemic patients that were correctly identified. False-positive (FP) cases defined as non-anemic patients that were incorrectly identified as anemic and false-negative (FN) cases were anemic patients that were incorrectly identified as non-anemic. A P value of less than 0.05 was considered significant. All statistical analyses were performed with SPSS software (version 21.0, IBM).

3 Results

3.1 Patient characteristics

The study cohort eventually contained 84 participants (43 men; 41 women,) aged 29.27±11.88 years (mean±standard deviation; range, 6–55 years) at the pre-transplant time point and 1-month post-transplant (n = 82; age, 29.32±12.02 years), 2-month (n = 72; age, 29.13±12.16 years), 3-month (n = 75; age, 28.93±11.97 years), 6-month (n = 74; age, 29.64±11.34 years), and 12-month (n = 70; age, 29.65±11.77 years), respectively. The multiple comparisons analyzed by ANOVA demonstrated that the CT value of IVS-LV, LV/IVS and hemoglobin were significantly different among all subgroups (all P < 0.001 as shown in Table 1).

Table 1
The clinical and imaging characteristics of AA patients at pre- and post-transplant time points

Groups No. of Female (%) Age(years) LV (HU) IVS (HU) IVS-LV LV/IVS Hb (g/L)

Pre(n = 84) 41/84 (48.8%) 29.27±11.88 32.94±6.95 48.42±6.60 15.48±4.86 0.68±0.10 63.15±10.38

Post-1 m(n = 82) 40/82 (48.7%) 29.32±12.02 33.72±8.30 48.10±6.76 14.38±6.95 0.70±0.13 75.77±14.07

Post-2 m(n = 72) 37/72 (51.3%) 29.13±12.16 37.54±7.72 48.47±6.51 10.93±4.74 0.77±0.11 99.24±20.21

Post-3 m(n = 75) 34/75 (45.3%) 28.93±11.97 37.81±6.98 47.72±5.81 9.91±5.09 0.79±0.11 104.39±25.00

Post-6 m(n = 74) 36/74 (48.6%) 29.64±11.34 38.75±8.19 48.18±7.23 9.42±5.42 0.80±0.11 107.91±20.77

Post-12 m(n = 70) 33/70 (47.1%) 29.65±11.77 40.04±7.88 48.36±7.47 8.31±5.62 0.83±0.12 113.43±26.45

P Value 0.999 <0.001* 0.986 <0.001* <0.001* <0.001*

Groups	No. of Female (%)	Age(years)	LV (HU)	IVS (HU)	IVS-LV	LV/IVS	Hb (g/L)
Pre(n = 84)	41/84 (48.8%)	29.27±11.88	32.94±6.95	48.42±6.60	15.48±4.86	0.68±0.10	63.15±10.38
Post-1 m(n = 82)	40/82 (48.7%)	29.32±12.02	33.72±8.30	48.10±6.76	14.38±6.95	0.70±0.13	75.77±14.07
Post-2 m(n = 72)	37/72 (51.3%)	29.13±12.16	37.54±7.72	48.47±6.51	10.93±4.74	0.77±0.11	99.24±20.21
Post-3 m(n = 75)	34/75 (45.3%)	28.93±11.97	37.81±6.98	47.72±5.81	9.91±5.09	0.79±0.11	104.39±25.00
Post-6 m(n = 74)	36/74 (48.6%)	29.64±11.34	38.75±8.19	48.18±7.23	9.42±5.42	0.80±0.11	107.91±20.77
Post-12 m(n = 70)	33/70 (47.1%)	29.65±11.77	40.04±7.88	48.36±7.47	8.31±5.62	0.83±0.12	113.43±26.45
P Value		0.999	<0.001*	0.986	<0.001*	<0.001*	<0.001*

Note: All data were means±standard deviations; LV: left ventricular; Hb: hemoglobin; IVS: interventricular septum; IVS-LV: difference between interventricular septum and left ventricular cavity; LV/IVS: ratio between left ventricular cavity and interventricular septum. *P < 0.05 the difference was statistically significant in comparison with all subgroups.

3.2 Reliability of the measurements and correlation of the CT value of IVS-LV, LV/IVS and hemoglobin

The parameter measurements of CT value of LV cavity and interventricular septum demonstrated excellent inter-observer reliability (ICC: 0.923, 95% CI: 0.899, 0.941, P < 0.001).The scatter plots of CT value for IVS-LV, LV/IVS and hemoglobin level were nearly linear relationship (Fig. 3). There was a negative linear correlation between the CT value of IVS-LV and hemoglobin (r = –0.618, P < 0.001). The CT value of LV/IVS was positively correlated with hemoglobin (r = 0.628; P < 0.001). The correlations of CT value of IVST-LV and LV/IVST with hemoglobin were similar, but the correlation coefficient of LV/IVST was slightly higher.

Fig. 3

Scatter plot and simple linear regression analysis. (A) The correlation of CT value of IVS-LV and hemoglobin. The correlation coefficient was–0.618 (P < 0.001). (B) The correlation of CT value of LV/IVS and hemoglobin. The correlation coefficient was 0.628 (P < 0.001). IVS-LV: difference between interventricular septum and left ventricular cavity; LV/IVS: ratio between left ventricular cavity and interventricular septum; Hb: hemoglobin.

3.3 The changes of IVS-LV, LV/IVS and serum hemoglobin at the pre- and post-transplant time points

The hemoglobin level increased continuously from pre-transplant time point to 12 months post-transplantation (Fig. 4A). The increased change of hemoglobin level had significant difference during the first month post-transplantation (P < 0.05), but they had no significant difference after 2 months post-transplantation (P > 0.05). Both the CT value of IVS-LV and LV/IVS had similarly variation tendency of hemoglobin with opposite directions in the pre- and post- transplant time points (Fig. 4B, 4C).

Fig. 4

The variation tendency of different indexes for AA patients during the treatment of a-HSCT. (A) Serum hemoglobin. (B) CT value of IVS-LV. (C) CT value of LV/IVS. Month 0: pre-transplant (within 2 weeks before transplantation); Hb: hemoglobin; IVS-LV: difference between interventricular septum and left ventricular cavity; LV/IVS: ratio between left ventricular cavity and interventricular septum.

3.4 ROC curve analysis

The area under the curve (AUC) of ROC analysis of the CT value of IVS-LV and LV/IVS in diagnosing anemia was 0.816 (95% CI: 0.777, 0.855; P < 0.001) and 0.814 (95% CI: 0.771, 0.849; P < 0.001), respectively (Fig. 5). According to the Youden index, when the CT value of the LV-IVS threshold was set at 11.5HU for diagnosing anemia, the sensitivity was 74.7% and the specificity reached 73.8%. When the CT value of LV/IVS was 0.77, the sensitivity and specificity for diagnosing anemia were 77.4% and 70.4%, respectively. The cases of TP, TN, FP and FN assessed by the CT threshold value of IVS-LV and LV/IVS in diagnosing anemia were showed in confusion matrix (Fig. 6).

Fig. 5

Comparison of receiver operating characteristic (ROC) curves of the CT value of IVS-LV and LV/IVS in diagnosing anemia. IVS-LV: difference between interventricular septum and left ventricular cavity; LV/IVS: ratio between left ventricular cavity and interventricular septum.

Fig. 6

The confusion matrix showed the cases of TP, FP, TN and FN in diagnosis of anemia by using the CT threshold value of IVS-LV(A), LV/IVS(B) of 11.5HU, 0.77, respectively. IVS-LV: difference between interventricular septum and left ventricular cavity; LV/IVS: ratio between left ventricular cavity and interventricular septum. TP: true-positive, FP: false-positive, TN: true-negative, FN: false-negative. 1: anemia, 0: non-anemia.

4 Discussion

Our study systematically focused on the accuracy of CT value of IVS-LV, LV/IVS in diagnosis of anemia for AA patients undergoing a-HSCT. The most important findings of the study were as follows: i) The CT value of LV/IVS was also correlated with serum hemoglobin, which was similar to that of IVS-LV. ii) The variation tendency of CT value of IVS-LV and LV/IVS was consistent with hemoglobin for AA patients in one year follow up of a-HSCT. iii) ROC analyses confirmed good sensitivity and specificity of the CT value of IVS-LV and LV/IVS to diagnose anemia.

AA patients usually had pulmonary infections and noninfectious complication, include pulmonary edema, engraftment syndrome, alveolar hemorrhage, et al. during the treatment of a-HSCT, which were the major causes of morbidity and mortality [14 –16]. Chest CT played an important role to diagnosis pulmonary disease, which provides an opportunity to quantify the CT attenuation of LV cavity and interventricular septum in the same field of view. Anemia as one of the noteworthy clinical symptoms has significant influence on patient’s quality of life and prognosis of disease.

Previous studies had suggested that simple attenuation measurements in CT scans correlated with hemoglobin level, which may be useful in predicting anemia [6 , 17–20]. Doppman et al. [18] first proposed that patients with severe anemia had a low-density LV blood pool on CT scans, which could be discerned by naked eyes. Title et al. [7] demonstrated that objective measurement of CT values in LV cavity performed significantly better accuracy and reliability than subjective assessment in differentiating an anemic from nonanemic state. Our study demonstrated that both the CT value of IVS-LV and LV/IVS had accuracy diagnosis of anemia for AA patients. Both of those values had well correlated with serum hemoglobin, though AA patients suffered with peripheric pancytopenia, including decreased hemoglobin, platelet and neutrophil counts. It was consistent with the previous research. Norman et al. [21] made twelve blood samples with varying hemoglobin concentrations by separating the cells of freshly heparinized blood from plasma and then the curves were constructed to relating the attenuation values to varying concentrations of hemoglobin. Finally, they found that the CT attenuation coefficient for whole blood was related linearly to serum hemoglobin, but less correlated with other cellular components, such as platelets and neutrophils. The attenuation values were predominantly due to the iron content of the hemoglobin molecule. Del Marmol et al. [22] also reported that the CT attenuation of in vitro blood primarily relates to the concentration of erythrocytes.

It had been suggested that measurements of the relative value of difference in CT attenuation between LV cavity and interventricular septum was better than the absolute value of LV cavity in diagnosing anemia [9, 11]. As we known, CT attenuation value itself is inaccurate and nonuniformity, because of X-ray fan, maladjustment, detector asymmetry, short-and long-term drift [7], which may be eliminated by the relative value of IVS-LV. In our study, we found that CT value of LV/IVS had well correlated with hemoglobin with a mildly higher correlation coefficient than the value of IVS-LV. The ratio value with higher sensitivity was expected to effectively reduce the offset artefacts, noise, or other uncertainties involved in image acquisition, reconstruction, which may enhance the accuracy of diagnosis. Furthermore, we found that the variation tendency of CT attenuation of IVS-LV and LV/IVS was consistent with hemoglobin in one-year follow-up period for AA patients undergoing a-HSCT. With the increasing of serum hemoglobin after HSCT, the CT attenuation of the LV cavity gradually increased. Since the density of interventricular septum composed of myocardial tissue changed less in the exclusion of endocardial disease. Therefore, the difference CT value of the LV cavity and interventricular septum decreased, and the ratio value increased.

Though the CT value of IVS-LV and LV/IVS had good correlation with hemoglobin, but they didn’t demonstrate the ability to diagnosing anemia. Since correlation is not a complete description of two-variable data, even they had linear relationship. Besides the correlation, we may need to rethink a more complete description of the data to make the conclusions. In this regard, we added the ROC analysis to assess accuracy in diagnosing anemia. The ROC analysis indicated that both the CT value of IVS-LV and LV/IVS had good effects in diagnosing anemia for AA patients (AUC of 0.816 and 0.814, respectively). When the threshold of CT attenuation of IVS-LV and LV/IVS was set at 11.5 HU and 0.77, respectively, both the sensitivity and specificity in diagnosing anemia were good and similar. In the previous studies, patients usually were mild to moderate anemia, and fewer patients had severe anemia. Besides, the causes of anemia were different, including chronic nephritis, renal failure, leukemia and malignant tumors, et al. but not the hypoplasia of bone marrow. Gennari et al. [23] demonstrated that the low-dose CT, as a part of a cardiac SPECT exam, may allow for identification of anemic patients. The optimal threshold to differentiate anemic from non-anemic patients was the difference CT value of LV cavity and interventricular septum was 4.5 HU. In fact, there were only 7 patients suffered with severe anemia (Hb < 60 g/L) among the 229 anemic patients. The optimal threshold of 4.5 HU was significantly lower than our study.

Lan et al. [9] reported when the CT attenuation threshold of IVS-LV was set at 10HU to diagnose anemia, the specificity was 100%, but the sensitivity was only 32.7%. If the threshold of IVS-LV was set at 12 HU, though the positive predictive value could reach to 100%, but there was no false-negative case. They considered that the CT threshold of IVS-LV > 10HU in male or > 12HU in female may be an effective value to assess severe anemia. Zhou et al. [11] divided anemia patients into three groups according to the level of hemoglobin, when the CT value of IVS-LV was set at (4.5–8.5) HU in diagnosing moderate anemia, the sensitivity and specificity were 93.9% and 60.0%, and the difference CT value greater than 13.5HU to diagnose severe anemia, the sensitivity was 82.4% and the specificity was 84.6%, respectively. But it had limited value to diagnose mild anemia with AUC as 0.471 (P > 0.05).

In our study, most AA patients suffered with moderate to severe anemia in pre-transplant and the first 2 months after transplantation, with the average hemoglobin of 63–99 g/L. During the last few months (3–12) in one year follow up after transplantation, the ‘health’ HSCT recipients also had lower hemoglobin than the normal laboratory, with the average hemoglobin of 104–113 g/L, which confirmed the observations of Koljonen et al. [13]. When we defined the critical level of anemia as less than 90 g/L, those ‘health’ HSCT recipients could be defined as non-anemic. Thus, we set the CT threshold of IVS-LV of 11.5HU to differentiate anemic from non-anemic, which were especially useful for AA patients with moderate to severe anemia. Of note, The CT threshold was basically consistent with the previous research in diagnosing moderate to severe anemia.

Apart from the retrospective design, this study has several limitations. First, although it is easy to evaluate anemia by chest CT images, blood test remains the standard tool to assess serum hemoglobin. It may be the reason why the assessment of hemoglobin acquired from chest CT imaging has not been widely use in clinical so far. However, not every patient was available of blood test at the time point of imaging, especially for AA patients who were prone to bleeding or coagulation disorders and not suitable for repeatedly blood tests. Besides, it may provide an increasing opportunistic usage of CT data with the advent of artificial intelligence including more accurate or efficient segmentation of the regions of interest [24] and automatic estimation of hemoglobin using CT images [25]. Thus, it may be possibility to add complementary evaluation of hemoglobin to each radiology report in routine diagnosis. Second, we did not conduct a cohort study with different sexes for AA patients, because some transplanted matches of donors and recipients were of different genders during the process of HSCT. Further research was required to confirm the effect of different genders on the results.

5 Conclusions

In conclusion, the non-invasive chest CT can offer a new possibility to complementary evaluation of anemia for AA patients in a frequent clinical setting. Both the CT attenuation of LV/IVS and IVS-LV had similar accuracy in diagnosing anemia for AA patients. When the threshold of CT attenuation of IVS-LV and LV/IVS was set at 11.5HU and 0.77, respectively, it will be helpful for radiologists to differentiate an anemic from non-anemic state for AA patients during the treatment of HSCT.

Footnotes

Acknowledgments

This study was supported by research funds from the Science and Technology Planning Project of Guangzhou (No.202002030427), and Basic and Applied Basic Research Foundation of Guangdong Province (No.2021A1515011288).

References

Peslak

S.A.

, Olson

, Babushok

D.V.

Diagnosis and treatment of aplastic anemia, Curr Treat Options Oncol 18(12) (2017), 70.

Luzzatto

, Risitano

A.M.

Advances in understanding the pathogenesis of acquired aplastic anaemia, Br J Haematol 182(6) (2018), 758–776.

Kojima

Why is the incidence of aplastic anemia higher in Asia?

Expert Rev Hematol 10(4) (2017), 277–279.

Georges

G.E.

, Doney

, Storb

Severe aplastic anemia: allogeneic bone marrow transplantation as first-line treatment, Blood Adv 2(15) (2018), 2020–2028.

Chaudhry

A.A.

, Gul

, Chaudhry

et al., Quantitative evaluation of noncontrast computed tomography of the head for assessment of anemia, J Comput Assist Tomogr 39(6) (2015), 842–848.

Foster

, Nolan

R.L.

, Lam

Prediction of anemia on unenhanced computed tomography of the thorax, Can Assoc Radiol J 54(1) (2003), 26–30.

Title

R.S.

, Harper

, Nelson

et al., Observer performance in assessing anemia on thoracic CT, AJR Am J Roentgenol 185(5) (2005), 1240–44.

Kamel

E.M.

, Rizzo

, Duchosal

M.A.

et al., Radiological profile of anemia on unenhanced MDCT of the thorax, Eur Radiol 18(9) (2008), 1863–1868.

Lan

, Nishihara

, Nishitani

Accuracy of computed tomography attenuation measurements for diagnosing anemia, Jpn J Radiol 28(1) (2010), 53–57.

10.

Jung

, Groth

, Bley

T.A.

et al., Assessment of anemia during CT pulmonary angiography, Eur J Radiol 81(12) (2012), 4196–4202.

11.

Zhou

Q.Q.

, Yu

Y.S.

, Chen

Y.C.

et al., Optimal threshold for the diagnosis of anemia severity on unenhanced thoracic CT: A preliminary study, Eur J Radiol 108 (2018), 236–241.

12.

Killick

S.B.

, Bown

, Cavenagh

et al., Guidelines for the diagnosis and management of adult aplastic anaemia, Br J Haematol 172(2) (2016), 187–207.

13.

Koljonen

P.A.

, Lie

A.K.

, Liang

et al., Determinants of peripheral blood count value in long-term survivors of allogeneic hemopoietic stem cell transplantation, Hematol Oncol 27(2) (2009), 98–101.

14.

Lertpongpiroon

, Rattarittamrong

, Rattanathammethee

et al., Infections in patients with aplastic Anemia in Chiang Mai University, BMC Hematology 18 (2018), 35.

15.

Franquet

, Müller

N.L.

, Lee

K.S.

et al., High-resolution CT and pathologic findings of noninfectious pulmonary complications after hematopoietic stem cell transplantation, AJR Am J Roentgenol 184(2) (2005), 629–637.

16.

Kunihiro

, Tanaka

, Kawano

et al., High-resolution CT findings of pulmonary infections in patients with hematologic malignancy: comparison between patients with or without hematopoietic stem cell transplantation, Jpn J radiol 40(8) (2022), 791–799.

17.

Collins

A.J.

, Gillespie

, Kelly

B.E.

Can computed tomography identify patients with anaemia?, The Ulster Medical Journal 70(2) (2001), 116–118.

18.

Doppman

J.L.

, Rienmuller

, Lissner

The visualized interventricular septum on cardiac computed tomography: a clue to the presence of severe anemia, J Comput Assist Tomogr 5(2) (1981), 157–160.

19.

Duy

, Badeeb

, Duy

et al., CT attenuation of acute subdural hematomas in patients with anemia, J Neuroimaging 29(4) (2019), 536–539.

20.

Mahmoudi

, Martin

S.S.

, Ackermann

et al., Potential of high dimensional radiomic features to assess blood components in intraaortic vessels in non-contrast CT scans, BMC Med Imaging 21(1) (2021), 123.

21.

Norman

, Price

, Boyd

et al., Quantitative aspects of computed tomography of the blood and cerebrospinal fluid, Radiology 123(2) (1977), 335–338.

22.

Del Marmol

Coulier

, “Black” and “white” blood on unenhanced CT, J Belg Soc Radiol 103(1) (2019), 16.

23.

Gennari

A.G.

, Grünig

, Benz

D.C.

et al., Low-dose CT from myocardial perfusion SPECT/CT allows the detection of anemia in preoperative patients, J Nucl Cardiol (2022), 10.1007/s12350-021-02899-x.

24.

Tan

, Zhou

, Li

et al., Automated vessel segmentation in lung CT and CTA images via deep neural networks, J Xray Sci Technol 29(6) (2021), 1123–1137.

25.

Zopfs

, Rinneburger

, Pinto Dos Santos

et al., Evaluating anemia using contrast-enhanced spectral detector CT of the chest in a large cohort of 522 patients, Eur Radiol 31(6) (2021), 4350–4357.