

ORIGINAL ARTICLE 

Year : 2022  Volume
: 18
 Issue : 3  Page : 718724 

Prediction of chronic kidney disease in abdominal cancers radiation therapy using the functional assays of normal tissue complication probability models
Ameneh Haghbin^{1}, Ahmad Mostaar^{2}, Reza Paydar^{1}, Mohsen Bakhshandeh^{3}, Alireza Nikoofar^{4}, Mohammad Houshyari^{5}, Susan Cheraghi^{6}
^{1} Department of Radiation Sciences, Faculty of Allied Medicine, Iran University of Medical Sciences, Tehran, Iran ^{2} Department of Medical Physics and Biomedical Engineering, Shahid Beheshti University of Medical, Tehran, Iran ^{3} Department of Radiology Technology, School of Allied Medical Sciences, Shahid Beheshti University of Medical Sciences, Tehran, Iran ^{4} Department of Radiation Oncology, Faculty of Medicine, Iran University of Medical Sciences, Tehran, Iran ^{5} Department of Radiation Oncology, Faculty of Medicine, Shahid Beheshti University of Medical Sciences, Tehran, Iran ^{6} Department of Radiation Sciences, Faculty of Allied Medicine, Iran University of Medical Sciences; Radiation Biology Research Center, Iran University of Medical Sciences, Tehran, Iran
Date of Submission  30Jan2021 
Date of Acceptance  14Jul2021 
Date of Web Publication  25Jul2022 
Correspondence Address: Susan Cheraghi Department of Radiation Sciences, Faculty of Allied Medicine, Iran University of Medical Sciences, Radiation Biology Research Center, Iran University of Medical Sciences. Shahid Hemmat Highway, P.O. Box 1449614535, Tehran Iran
Source of Support: None, Conflict of Interest: None  Check 
DOI: 10.4103/jcrt.jcrt_179_21
Aim: The purpose of this study is to predict chronic kidney disease (CKD) in the radiotherapy of abdominal cancers by evaluating clinical and functional assays of normal tissue complication probability (NTCP) models. Materials and Methods: Radiation renal damage was analyzed in 50 patients with abdominal cancers 12 months after radiotherapy through a clinical estimated glomerular filtration rate (eGFR). According to the common terminology criteria for the scoring system of adverse events, Grade 2 CKD (eGFR ≤30–59 ml/min/1.73 m^{2}) was considered as the radiation therapy endpoint. Modeling and parameter estimation of NTCP models were performed for the Lymanequivalent uniform dose (EUD), the logitEUD critical volume (CV), the relative seriality, and the mean dose model. Results: The confidence interval of the fitted parameters was 95%. The parameter value of D_{50} was obtained 22–38 Gy, and the n and s parameters were equivalent to 0.006 –3 and 1, respectively. According to the Akaike's information criterion, the mean dose model predicts radiationinduced CKD more accurately than the other models. Conclusion: Although the renal medulla consists of many nephrons arranged in parallel, each nephron has a seriality architecture as renal functional subunits. Therefore, based on this principle and modeling results in this study, the whole kidney organs may have a serial–parallel combination or a secret architecture.
Keywords: Chronic kidney disease, normal tissue complication probability models, radiotherapy
How to cite this article: Haghbin A, Mostaar A, Paydar R, Bakhshandeh M, Nikoofar A, Houshyari M, Cheraghi S. Prediction of chronic kidney disease in abdominal cancers radiation therapy using the functional assays of normal tissue complication probability models. J Can Res Ther 2022;18:71824 
How to cite this URL: Haghbin A, Mostaar A, Paydar R, Bakhshandeh M, Nikoofar A, Houshyari M, Cheraghi S. Prediction of chronic kidney disease in abdominal cancers radiation therapy using the functional assays of normal tissue complication probability models. J Can Res Ther [serial online] 2022 [cited 2022 Aug 10];18:71824. Available from: https://www.cancerjournal.net/text.asp?2022/18/3/718/351812 
The renal failure caused by ionizing radiation was first observed in 1906 and more comprehensively described by Hall and Whipple in 1919 but quickly dismissed. Subsequently, radiationinduced renal insufficiency was described in human's abdomen irradiation for more than 100 years.^{[1]} Radiationinduced renal damage is classified according to the amount of the received radiation dose and dose fractionation and other systemic cancer treatments that patients receive.^{[2]}
The kidneys are lateresponding tissues and radiationinduced nephropathy can result from the kidneys' exposure during abdominal cancers. Radiation therapy (RT) can appear in months to years after exposure as acute or chronic kidney disease (CKD). This delay is associated with a slow turnover in the renal tissue compared to acute responding tissues, such as bone marrow or the gastrointestinal epithelial tissue.^{[3]}
Emami et al. described the probability of developing normal tissue complications and suggested organ tolerance limits based on the volume of the organ irradiated to various doses.^{[4]} Milano et al. updated Emami's normal tissue dose constraints.^{[5]} However, they were unable to modify renal dose tolerance limits given the insignificant availability of published data.
In the quantitative analyses of normal tissue effects in the clinic (QUANTEC), there have been previous studies reviewing the effects of partial kidney irradiation.^{[6]} The kidneys are doselimiting organs in the abdominal RT, and previous studies were not conducted based on the performance evaluation of normal tissue complication probability (NTCP) in RT planning subject to computed tomography (CT) scans images. While evaluating the treatment planning in RT, a plan is acceptable only if it leads to achieving a high tumor control probability at low NTCP. In recent years, several NTCP models have been developed to predict normal tissue damage following RT.^{[7],[8]} These models have been introduced as new indicators for the evaluation and optimization of RT plans and. Moreover, based on these models, the degree and the definition of radiation damage have been determined through scoring systems. These standard scoring systems quantify the effects of radiation based on the type of structural and functional findings. This study is aimed to introduce the best predictive radiobiological NTCP model to determine the CKD in RT of abdominal cancers.
> Materials and Methods   
Patient selection
The study includes prospective analysis in 50 patients aged 25–80 years with abdominal cancers (gastric 84%, cardiac 4%, gall bladder 6%, and pancreas 6%) treated with RT.
To participate in the study, patients should have had normal results on kidney functional tests prior to RT.^{[9]} Patients with kidney disease, cisplatin chemotherapy, diabetes, and hypertension were excluded from the study. Twelve months after the RT, urine and biochemical analyses were performed in all patients to evaluate CKD. CKD was defined as an endpoint through the estimated glomerular filtration rate (eGFR). Each patient was compared with him/herself according to prior RT eGFR. Kidney damages of Grade 2 (eGFR = 5930 ml/min/1.73 m^{2}) or above were considered to describe CKD based on the standard scoring systems^{[10]} [Table 1] as the common terminology criteria for adverse events (CTCAE version 5).  Table 1: The common terminology criteria for adverse events scoring system for chronic kidney disease definition
Click here to view 
The procedures were carried out under the ethical standards of the responsible committee on human experimentation. Patients gave informed consent before enrollment (IR.IUMS.REC 1396.31892).
Radiotherapy
Patients were treated with the external beam RT to the abdomen. The RT procedure was carried out using a total dose of 40–60 Gy by a 6 MV photon beam of a linear accelerator planned with an isocentric technique to deliver 1.82 Gy/d in 5 consecutive days per week. All patients were treated using a threedimensional treatment planning system (3DTPS) (CorePLAN) based on patients' CT scans images.
Planning the treatment volume and organs at risk were delineated by an experienced radiation oncologist. The right and left kidneys were delineated individually and combined as a composite total kidney volume.^{[11]} Total kidney differential dose–volume histograms (DVHs) were calculated and imported as input to the NTCP modeling program. Different physical doses were converted into biologically effective doses in 2 Gy fractions (BED2), and the a/b ratio for normal tissue was set to 3 Gy (Equation 1).
BED = D (1 + dα/[β]) (Equation 1)
Normal tissue complication probability models
Modeling the dose–response for kidney was performed using 5 NTCP models, including the Lymanequivalent uniform dose (LEUD), the logitEUD (LogEUD), the critical volume (CV), the relative seriality (RS), and the mean dose model. [Table 2] shows the examined models and descriptions of their parameters.  Table 2: Examined normal tissue complication probability models with a summary of relationship and definition of parameters
Click here to view 
If the EUD dose is delivered uniformly over the organ, it would have the same biological effect as the heterogeneous dose described by the DVH (Equation 2).
In this expression, ν_{i} is the fraction of the reference volume irradiated to dose D_{i} in Gy, and α is a unitless parameter for tumor or speciﬁc normal tissues, which describes the dose–volume effect. The LEUD model is often used as a mean dose model by fixing the n parameter to 1.^{[12]} The n parameter indicates the volume effect of the organ being assessed. The LogEUD model^{[13],[14]} uses the logit formula coupled with the generalized EUD reduction algorithm (Equation 2).
The RS model is developed to account for the functional organization of functional subunits (FSUs) in an organ more efficiently.^{[15]} The s parameter describes the degree of seriality (the s value varies from zero for parallel organs to higher for increasing seriality). The s parameter in this model is inversely proportional to n in the LEUD model.
Maximum Likelihood fitting and goodness of fit
The maximum likelihood method^{[8],[16]} was employed to determine the best estimation of parameters in NTCP modeling. The best estimation of the parameter values are those that maximize the natural logarithm of the likelihood function [ln L()] (Equation 4).
The vector parameter in these is composed of the models' parameters; NTCPi is used when the i^{th} patient experiences a side effect, and 1 – NTCPi is used when no side effect is observed.
The search for the best values in NTCP modeling parameters was performed by the global optimization toolbox in the MATLAB environment. The 95% confidence intervals of the fitted parameters were calculated using the profile likelihood method.
The goodness of fit was applied by the twosample Kolmogorov–Smirnov (KS) test to determine significant differences between the cumulative distribution functions. The KS was used to compare the observed (F1) and theoretical (F2) cumulative distribution functions. The test statistic D (Equation 5) detects the largest vertical difference between the empirical distribution and the hypothesized distribution functions.
D = max (F1(x)F2(x)) (Equation 5)
The null hypothesis stated that the observed and theoretical functions have the same continuous distribution. The test statistic (D) will be 1 if the test rejects the null hypothesis at the 5% probability level (P = 0).
The secondorder Akaike's information criterion (AIC) (Equation 6) was used for ranking the accepted models^{[8]} as follows:
AIC = −2LL + 2k + ([2k [k + 1]]/[[n – k − 1]]) [Equation 6]
LL is the natural logarithm of the fitted model's likelihood, k is the number of parameters, and n is the number of patients. Models with smaller values of the AIC are considered to provide a better fit to the data than models with larger values of this index.
> Results   
Patient and treatment plan characteristics
All 50 patients were assessed by laboratory eGFR and dosimetric data after the RT. A summary of the demographic, clinical, and dosimetric characteristics of the patients is shown in [Table 3].  Table 3: Patient demographics (n=50), values of their dose and volume of kidneys, and their chronic kidney disease damage
Click here to view 
The mediandelivered radiation dose to target (gross tumor volume, involved lymph nodes, and/or postoperative tumor bed) was 50.4 Gy.
Changes in eGFR were used to assess the renal function. Grade 2 of CKD (eGFR ≤30–59 ml/min/1.73 m^{2}) was considered as the endpoint. Twentynine of 50 patients (58%) classified in Grade 2, Grade 3, or 4 renal damage were not observed in our study.
Normal tissue complication probability models analysis
The resulting optimum parameters with 95% confidence intervals, the goodness of fit, and AIC values of the models are given in [Table 4]. According to the goodness of fit assessed by the twosample KS test, the CV NTCP model is rejected. The fitted value of the D50 ranged from 22 to 60 Gy. The fitted value of m was 0.45 in the LEUD model. [Figure 1] shows the NTCP curves for the LEUD, LogEUD, RS, and the mean dose model.  Table 4: Model parameters with 95% confidence interval, goodness of fit values and Akaike's information criterion of the models for chronic kidney disease endpoint
Click here to view 
 Figure 1: Experimental points and estimated normal tissue complication probability curves for kidneys endpoint by applying (a) Lyman equivalent uniform dose, (b) logit equivalent uniform dose, (c) relative seriality, (d) mean dose models (NTCP curve, ° Estimation of occurrence, *Experimental points)
Click here to view 
The AIC indicated that the mean dose model holds the highest ranking for CKD prediction. The LogEUD and LEUD models were the next best models, with the lowest AIC value.
> Discussion   
This study aimed to establish the dose–response relationships and derive NTCP parameters of radiationinduced CKD as an endpoint in patients with abdominal cancers. Due to the kidneys' sensitivity to radiation, they are considered doselimiting organs in RT. The tolerance dose, the latency period, and clinical toxicities secondary to renal irradiation were described by Kunkler et al.^{[17],[18]} Acute toxicity of kidney occurred in 6–12 months, late effects in 18 months following completion of RT, and development of symptomatic nephropathy appears longer latency after radiation.^{[19]}
Dewit et al. noted an inverse relationship between dose and latency period to developing late effects.^{[20]} Using renal tolerance of 20 Gy to both kidneys, Krochak and Baker described the clinical presentation and pathophysiology of radiationinduced renal dysfunction.^{[21]}
According to several studies in this field, an acceptable tolerance dose has not been determined for the kidneys. Emami et al. analyzed radiationinduced kidney damage and the possibility of normal tissue complications through RT data. Their study stated that the irradiated organ's tolerance depends on the volume of the irradiated organ with different radiation doses. In Emami's study, the kidneys' tolerance for a 5% probability of complication in 5 years was 23 Gy for the whole organ, 30 Gy for twothird volume, and 50 Gy for onethird volume.^{[22]}
Milano et al. updated the dose limits of natural tissue described by Emami.^{[5]} However, they were not able to change the renal dose tolerance. The challenges were over in diagnosing late clinical effects and the definition of renal dose limits.
Johnson et al. observed an exponential decrease in renal function in patients with gastric cancer after postoperative RT.^{[23]}
The maximum dose, mean dose, minimum dose, isodose, and DVH are usually used to evaluate the treatment plan, but this information does not consider the tissues under the study's biological properties. NTCP models show appropriate behavior in describing the experimental data concerning biological properties. These models use DVHs as input to calculate the probability of normal tissue complication.
Withers shown in integral response architectures such as kidneys, complication will occur when a substantial fraction of the FSUs is damaged.^{[24]} In 1991, Niemierko and Goitein investigated the role of organ structure and architecture in the NTCP model based on dose–volume algorithms and reported that normal tissue responds differently to radiation due to its architecture.^{[25]} According to their findings, some organs, such as the kidneys, liver and lung had an integral response so complication occurs when a significant portion of the FSU is damaged. In 1993, these two researchers introduced the CV model and examined the model behavior for the kidneys, considering clinical nephritis as a functional endpoint. They stated that the kidney is a parallel organ because nephrons like FSUs have a parallel relationship. Based on the CV model of Niemierko and Goitein and the description by Kallman, FSU may be specified for different regions of the kidney.^{[26]}
Yorke et al. assumed the kidneys as parallel organs in studying the possibility of complications due to radiation. The linearquadratic model is used to gain the dose–response of individual cells within an FSU. This model's predictions are compared with an empirical powerlaw function for uniform partial and whole organ irradiation. The complication is produced if a sufficiently large fraction of the FSUs is inactivated by radiation, and an FSU is inactivated only when all the clonogenic cells within it are killed.^{[27]} The other study is based on the normal tissue dose–volume tolerance in QUANTEC, where the TD5 and TD50 of whole kidneys are 15–18 Gy and 28 Gy, respectively.^{[6]} In the same study, the kidneys were introduced as an example of parallel organs in which the damage to a subunit does not affect the other subunits' function. The kidneys are likened to LKB models with n ≈ 1 like as a parallel organ.^{[28]}
In 2014, Chang et al. introduced kidneys as parallel structurally defined with a group of FSUs (nephrons). In a parallel organ, the loss of function in one part of the organ only affects that part of the organ. There is a threshold volume effect. The partial organ effect does not always correlate with the whole organ function.^{[29]} Kallman et al., reported that the parenchyma of kidney consists of parallel nephrons and each nephron consists of a series of FSUs, all with several regenerative units arranged in parallel. They assume schematic tissue organization structures in the kidney as a parallel–serial model. Each nephron consists of a series of FSUs, all with several regenerative units arranged in parallel. They assume schematic tissue organization structures in the kidney as a parallel–serial model [Figure 2]. The first parallel structure is the capillary system inside the glomerular capsule, followed by the capsule itself and the limbs and Henle's loop. These are the FSUs of a kidney as described by equations.^{[15]}
Olsen et al., in examining model parameters based on clinical data and calculating the NTCP for whole or part of an organ, stated that the kidney is a parallel organ while it is possible nephrons can be interpreted as a functional unit in their secret structure.^{[30]}
Wessels et al., in a study of the effect of model hypotheses on renal dosimetry and responseoutcome therapy for radionuclide therapy with the MCNP Monte Carlo code, stated that nephrons are essential renal FSUs and they are responsible for the formation of urine and extends from the cortex to the center (medulla) of the kidney.^{[31]}
Rubin et al. stated that the architecture of kidneys could be defined as “parallel” in terms of the arrangement of FSUs (nephrons) but with some “series” functions.^{[32]} Some organs present a combination of serial and parallel FSU arrangements. The lung airways are arranged in serial, but the smaller airways have a parallel FSU setup. Although the brain is arranged in a parallel FSU structure, the loss of a critical brain region such as optic chiasm can lead to serious complications.
Kim et al. in 2017, in their studies on the factors, predicted to improve damage in different types of organs based on their architecture and the initial state of the injury healing mechanism, renal hilum, and vascular trunk in the category of tissues with secret architecture.^{[33]} In our study, the architectural models did not predict CKD. Furthermore, the AIC showed the highest ranking for the mean dose model. The LogEUD and LEUD models were the next best model, with the lowest AIC value. However, this study is open and can be done with a larger sample size. Our study results confirm Kallman and Kim DN's study that the whole organ of the kidney is a secret or combination architecture (seriesparallel), which can be the cause of inconsistencies in existing studies. However, the significant problems in using existing mathematical models to estimate NTCP are needed further studies.^{[6]}
Since the number of active nephrons changes after receiving radiation dose to irradiated volume, as CKD progresses, the number of functioning nephrons decreases. With the eGFR of approximately 60 mL/min/m^{2}, a critical loss of function occurs with no return to recovery.^{[34]} The remaining nephrons function in an increased compensatory mode with hyperfiltration, and the degree of possible compensation is determined by the renal functional reserve.^{[35]}
According to our experience in this study and the results obtained by some studies, although each kidney comprises many nephrons (glomeruli and tubes) arranged in parallel, each nephron has a seriality structure. In addition, it is suggested that the structure type of cortex and renal pelvis structure should be examined furthermore.
> Conclusion   
Our study showed that the phenomenological NTCP models can predict CKD better than the models based on FSU models, such as RS and CV models. Although the renal medulla comprises many nephrons arranged in parallel, each nephron, as a renal FSU, has a seriality architecture whose radiation damage is irreversible. Therefore, based on this principle and modeling results in this study, the kidney's whole organ may be a combination of “serialparallel” or secret architecture. Therefore, we need improved and more accurate models for such organs with a dependence beyond dose–volume relationships and tissue architecture. Finally, we conclude in organs such as kidneys, the pathophysiology of radiation injury is poorly understood and needs further study.
Financial support and sponsorship
This work was supported financially by the Research Chancellor of Iran University of Medical Sciences (96033131892).
Conflicts of interest
There are no conflicts of interest.
> References   
1.  Edsall DL. The attitude of the clinician in regard to exposing patients to the Xray. JamaJ Am Med Assoc 1906;47:14259. 
2.  Podoll AS, Amsbaugh MJ. Radiationassociated kidney injury. In: Renal Disease in Cancer Patients. Academic Press: Elsevier; 2014. p. 10513. 
3.  Hall E, Giacci A. Doseresponse relationships for model normal tissues. In: Radiobiology for the Radiologist. Wolters Kluwer, Eight edition, 2019. p. 6567. 
4.  Brown AP, Urie MM, Barest G, Cheng E, Coia L, Emami BN, et al. Threedimensional photon treatment planning for Hodgkin's disease. Int J Radiat Oncol Biol Phys 1991;21:20515. 
5.  Milano MT, Constine LS, Okunieff P. Normal tissue tolerance dose metrics for radiation therapy of major organs. Semin Radiat Oncol 2007;17:13140. 
6.  Bentzen SM, Constine LS, Deasy JO, Eisbruch A, Jackson A, Marks LB, et al. Quantitative analyses of normal tissue effects in the clinic (QUANTEC): An introduction to the scientific issues. Int J Radiat Oncol Biol Phys 2010;76:S39. 
7.  Cheraghi S, Nikoofar A, Bakhshandeh M, Khoei S, Farahani S, Abdollahi H, et al. Normal tissue complication probability modeling of radiationinduced sensorineural hearing loss after headandneck radiation therapy. Int J Radiat Oncol Biol Phys 2017;93:132733. 
8.  Bakhshandeh M, Hashemi B, Mahdavi SR, Nikoofar A, Vasheghani M, Kazemnejad A. Normal tissue complication probability modeling of radiationinduced hypothyroidism after headandneck radiation therapy. Int J Radiat Oncol Biol Phys 2013;85:51421. 
9.  Ashcroft RE. The declaration of Helsinki. In: The Oxford Textbook of Clinical Research Ethics. Oxford University Press New York; 2008. p. 1418. 
10.  Diane MF Savarese, MD. Common Terminology Criteria for Adverse Events: (CTCAE). Cancer Therapy Evaluation Program; National Cancer Institute (US); 2013. 
11.  May KS, Khushalani NI, Chandrasekhar R, Wilding GE, Iyer RV, Ma WW, et al. Analysis of clinical and dosimetric factors associated with change in renal function in patients with gastrointestinal malignancies after chemoradiation to the abdomen. Int J Radiat Oncol Biol Phys 2010;76:11938. 
12.  Lyman JT. Complication probability as assessed from dosevolume histograms. Radiat Res 1985;104:S139. 
13.  Niemierko A. A generalized concept of equivalent uniform dose (EUD). Med Phys 1999;26:1100. 
14.  Lax I, Karlsson B. Prediction of complications in gamma knife radiosurgery of arteriovenous malformation. Acta Oncol 1996;35:4955. 
15.  Källman P, Agren A, Brahme A. Tumour and normal tissue responses to fractionated nonuniform dose delivery. Int J Radiat Biol 1992;62:24962. 
16.  Rancati T, Fiorino C, Sanguineti G. NTCP modeling of subacute/late laryngeal edema scored by fiberoptic examination. Int J Radiat Oncol Biol Phys 2009;75:91523. 
17.  Kunkler PB, Farr RF, Luxton RW. The limit of renal tolerance to xrays; an investigation into renal damage occurring following the treatment of tumours of the testis by abdominal baths. Br J Radiol 1952;25:192201. 
18.  Glezerman IG. Radiation nephropathy. Onconephrology, Springer; 2015. p. 2419. 
19.  Thompson PL, Mackay IR, Robson GS, Wall AJ. Late radiation nephritis after gastric Xirradiation for peptic ulcer. Q J Med 1971;40:14557. 
20.  Dewit L, Anninga JK, Hoefnagel CA, Nooijen WJ. Radiation injury in the human kidney: A prospective analysis using specific scintigraphic and biochemical endpoints. Int J Radiat Oncol Biol Phys 1990;19:97783. 
21.  Krochak RJ, Baker DG. Radiation nephritis clinical manifestations and pathophysiologic mechanisms. Urology 1986;27:38993. 
22.  Emami B, Lyman J, Brown A, Coia L, Goitein M, Munzenrider JE, et al. Tolerance of normal tissue to therapeutic irradiation. Int J Radiat Oncol Biol Phys 1991;21:10922. 
23.  Jansen EP, Saunders MP, Boot H, Oppedijk V, Dubbelman R, Porritt B, et al. Prospective study on late renal toxicity following postoperative chemoradiotherapy in gastric cancer. Int J Radiat Oncol Biol Phys 2007;67:7815. 
24.  Withers HR, Taylor JM, Maciejewski B. Treatment volume and tissue tolerance. Int J Radiat Oncol Biol Phys 1988;14:7519. 
25.  Niemierko A, Goitein M. Calculation of normal tissue complication probability and dosevolume histogram reduction schemes for tissues with a critical element architecture. Radiother Oncol 1991;20:16676. 
26.  Niemierko A, Goitein M. Modeling of normal tissue response to radiation: The critical volume model. Int J Radiat Oncol Biol Phys 1993;25:13545. 
27.  Yorke ED, Kutcher GJ, Jackson A, Ling CC. Probability of radiationinduced complications in normal tissues with parallel architecture under conditions of uniform whole or partial organ irradiation. Radiother Oncol 1993;26:22637. 
28.  Marks LB, Yorke ED, Jackson A, Haken RK, Constine LS, Eisbruch A, et al. Use of normal tissue complication probability models in the clinic. Int J Radiat Oncol Biol Phys 2010;76:S109. 
29.  David S. Chang Foster D. LasleyIndra J. DasMarc S. MendoncaJoseph R. Dynlacht. Basic Radiotherapy Physics and Biology, Normal tissue radiation responses. Basic Radiother Phys Biol, Springer 2014; 26575. 
30.  Olsen D, Kambestad BK, Kristoffersen DT. Calulation of radiation induced complication probabilities for brain, liver and kidney, and the use of a reliability model to estimate critical volume fractions. Br J Radiol 1994;67:121825. 
31.  Wessels BW, Konijnenberg MW, Dale RG, Breitz HB, Cremonesi M, Meredith RF, et al. MIRD pamphlet No. 20: The effect of model assumptions on kidney dosimetry and response – Implications for radionuclide therapy. J Nucl Med 2008;49:188499. 
32.  Rubin P, Constine LS, Marks LB. ALERTAdverse Late Effects of Cancer Treatment. Normal Tissue Specific Sites and Systems, Springer; 2014;2. 
33.  Kim DN, Medin PM, Timmerman RD. Emphasis on repair, not just avoidance of injury, facilitates prudent stereotactic ablative radiotherapy. Semin Radiat Oncol 2017;27:37892. 
34.  Chnaper HW. Remnant nephron physiology and the progression of chronic kidney disease. Pediatr Nephrol 2014;29:193202. 
35.  Ivanov D, Gozhenko A, Savitskaya L. Renoprotection and its association with eGFR and renal functional reserve. Kidneys 2018;7:23844. 
[Figure 1], [Figure 2]
[Table 1], [Table 2], [Table 3], [Table 4]
