Вы находитесь на странице: 1из 5

Computerized Medical Imaging and Graphics 54 (2016) 1–5

Contents lists available at ScienceDirect

Computerized Medical Imaging and Graphics


journal homepage: www.elsevier.com/locate/compmedimag

Evaluation of 3D-CRT, IMRT and VMAT radiotherapy plans for left


breast cancer based on clinical dosimetric study
Haiyun Liu a,1 , Xinde Chen b,1 , Zhijian He b,∗ , Jun Li c
a
Jiangxi University of Traditional Chinese Medicine, Nanchang, 330029, Jiangxi, China
b
Tumor Hospital of Jiangxi Province, Nanchang, 330029, Jiangxi, China
c
Jiangsu Subei People’s Hospital, Yangzhou, 225001, Jiangsu, China

a r t i c l e i n f o a b s t r a c t

Article history: Objective: This paper aims to compare dosimetric differences based on three types of radiotherapy
Received 15 June 2016 plans for postoperative left breast cancer. In particular, based on a clinical dosimetric study, the three-
Received in revised form dimensional conformal radiotherapy (3D-CRT), intensity- modulated radiation therapy (IMRT) and VMAT
20 September 2016
plans were implemented on 15 cases of postoperative patients with left breast cancer with prescription
Accepted 10 October 2016
doses of 5000 cGy.
Methods and results: Dose volume histogram (DVH) was used to analyze each evaluation index of clinical
Keywords:
target volume (CTV) and organs at risk (OARs). Except for homogeneous index (HI), D2 , each CTV eval-
Left breast cancer
3D-CRT
uation index of 3D-CRT plan was inferior to IMRT and VMAT plans (P < 0.05). Compared with the VMAT
IMRT plans, IMRT has a statistical significance only in Dmean , V95 (P < 0.05). On the contrary, Dmean pertaining
VMAT to the VMAT plan is much closer to the prescription dose with a V95 coverage rate as high as 97.44%. For
Dosimetry the infected lung, V5 , V10 of 3D-CRT were the lowest (P < 0.05), while V20 , V30 were the highest (P < 0.05)
among the three types of plans. Here, the V5 , V10 of infected lung were slightly higher (P < 0.05) for
the VMAT and IMRT plans. Each evaluation index of the contralateral lung and heart in 3D-CRT was
the lowest (P < 0.05). D1 of contralateral breast was lower in both IMRT and VMAT plans, which were
1770.89 ± 121.16 cGy and 1839.92 ± 92.77 cGy, respectively. While D1 of the spinal cord in IMRT and
VMAT plans was higher, which were 1990.12 ± 61.52 cGy and 1927.38 ± 43.67 cGy, respectively. When
the radiation dose of 500–1500 cGy was delivered to the normal tissues, 3D-CRT significantly shows the
lowest volume, VMAT is relatively higher. Monitor Units (MU) and treatment time (T) of VMAT were the
least, only 49.33% and 55.86% of those of IMRT.
Conclusion: The three types of plans can meet the clinical dosimetry demands of postoperative radiother-
apy for left breast cancer. The target of IMRT and VMAT plans has a better conformity, and the VMAT plan
takes the advantages of less MU and treatment time.
© 2016 Elsevier Ltd. All rights reserved.

1. Introduction missing the target area ought to be taken as a basic requirement for
radiotherapy (Palma et al., 2008).
Radiation therapy has become one of the vital measurements of Conventionally, the 3D-CRT has been one of the typical main
postoperative breast cancer treatment. It is also the most impor- plan solutions. Our methods are based on breast tangential field
tant means of improving the local control rate of tumor, as well as irradiation and lymphatic draining region irradiation by using cen-
to reduce complication of the normal tissues. The scope of radio- ter irradiation method of 1/4 tangent field and semi field. However,
therapy covers mainly the chest wall and lymph nodes at positions this comes with the disadvantages of poor target area fitness
above and under the collar bone. The radiotherapy technology is and higher complexity in our radiotherapy practice. In addition,
more complex, and in order to avoid overlap and omission of adja- IMRT being used for patients with post-operative breast cancer has
cent radiation, reducing radiation damage to normal tissue without become more and more common. However, because of the com-
plexity of the design of treatment plan, the field angle also has
important effects on the quality of treatment plan (Vicini et al.,
∗ Corresponding author. 2000). And the extensive treatment time would cause reduction
E-mail address: zhijianhe33@sina.com (Z. He). of relative biological effects, postural changes of patients during
1
These authors contributed equally to the manuscript. treatment, decrease of treatment accuracy and a series of problems.

http://dx.doi.org/10.1016/j.compmedimag.2016.10.001
0895-6111/© 2016 Elsevier Ltd. All rights reserved.
2 H. Liu et al. / Computerized Medical Imaging and Graphics 54 (2016) 1–5

With the rapid development of hardware and software in the field as starting and ending angle each way, with collimator
field of radiotherapy, a new type of technology known as the Volu- angle 5◦ , treatment couch angle 0◦ , maximum dose rate 600
metric Modulated Arc Therapy (VMAT), which combines the fixed MU/min, and algorithm model is AAA. In the process of design-
field IMRT and pull arc irradiation, has been developed recently. ing, doses of cold and hot points were optimized and adjusted
Dosimetry studies show that VMAT can be better than the fixed by defining dose shaping structure (DSS) (Otto, 2008).
field IMRT. The VMAT technology that is provided by Varian Eclipse
8.6 planning system has already been confirmed as VMAT mode 2.4. Plan assessment
by Otto, etc (Otto, 2008). To achieve rotation IMRT treatment, this
technology was use to calculate the state of multi-leaf collimator (a) We aim to evaluate the parameters including Dmean , D2 , D98 ,
motion by adopting reverse optimization algorithm, meanwhile by V90 , V95 , CI and HI for CTV. Herein, CI (Anon, 2010; Yin et al., 2011)
optimizing the dose rate and rotate speed of the rack. Due to the is expressed by
arc structure of the breast cancer, three radiotherapy techniques, Vt,ref Vt,ref
i.e., 3D-CRT, IMRT and VMAT, have been widely applied in clinic CI = × , (1)
Vt Vref
nowadays. The purpose of this paper was to compare the dosi-
metric parameters and to obtain the most superior radiotherapy where Vt stands for the target volume, Vt,ref stands for the target
technique, i.e., to obtain the optimal dosimetric distribution of the volume surrounded by reference isodose surface, Vref is the volume
target and to maximum reduce the dose delivered to the lung by of all areas surrounded by reference isodose surface. Here, CI ranges
designing three radiotherapy plans for a certain case. from 0 to 1, and higher CI values indicate better conformity. HI (Yin
et al., 2011) is given by
2. Materials and methods D2 − D98
HI = × 1000⁄0, (2)
Dprescription
2.1. Case selection
where D2 and D98 (dose received by the 2% and 98% of the vol-
We randomly selected 15 cases of postoperative radiotherapy ume, respectively) are metrics for minimum and maximum doses.
for breast cancer in our hospital during March 2011 and March Dprescription is the prescription dose, and lower HI values indicate
2013. The patients are all female, aged 38 to 65 years old, the aver- superior dose homogeneity of the target volume.
age age of 45 years old, and the primary lesions are left breast. (b) Clinical constraints for OARs: V5 , V10 , V20 , V30 and Dmean for
the infected lung; Dmean, V5 and V10 for the contralateral lung; Dmean
2.2. CT simulation location and target area delineation and and D1 for the contralateral breast; Dmean and V10 for the heart; D1
definition for the spinal cord.
All the data in this article utilized the SPSS15.0 for statistical
Patients raised both arms above their heads, and were fixed by analysis. We adopted the paired-samples T-test to perform the
the vacuum negative pressure bag. CT scans with a slice thickness comparison of dosimetry differences among 3 plans, which is based
of 5 mm were obtained using large aperture 16 rows spiral CT of GE on as the statistical difference (P < 0.05).
Medical Systems. CT scans ranges from the mandible to the thorax,
which completely cover all the adjacent normal tissues and organs 3. Results
such as lung, heart, opposite breast and the spinal cord, etc. The
clinical target volume (CTV), including the whole ipsilateral chest The following results are based on our experimentations:
wall and lymph node region around collar bone, was outlined by
the oncologist by using Varian eclipse 8.6 treatment planning sys- 3.1. Distribution of the target dose
tem (TPS), and the organs at risk (OARs) including ipsilateral lung,
contralateral lung, contralateral breast, heart, and the spinal cord As shown in Table 1, except for HI, D2 , all the evaluation param-
were delineated then. eters for CTV in 3D-CRT plan were inferior to those in IMRT and
VMAT plans (P < 0.05). However, CI was lowest in 3D-CRT plan and
2.3. Plan design also reflected in Fig. 1. Compared with VMAT plans, IMRT had a
statistical significance only for Dmean , V95 . Then, Dmean was much
The prescription dose given was 5000cGy, which was irradi- closer to the prescription dose, and V95 reached to 97.44% in VMAT
ated for 25 times, herein, for fractionated dose of 200cGy, 99% plan.
of CTV is supposed to receive at least 95% of prescription dose
(4750cGy). Clinical constraints are as follows: CTV ≤ 107%, the min- 3.2. Comparison of OARs dosimetry
imum dose ≥ 95%, V20 < 30% and the average dose Dmean < 1500 cGy
for the ipsilateral lung, the maximum dose of the spinal cord The results of OARs were shown in Table 1. For the ipsilateral
Dmax < 4000 cGy, V40 < 50% for heart and the dose delivered to the lung, V5 , V10 in 3D-CRT were the lowest (P < 0.05), while V20 , V30
contralateral lung and the contralateral breast should be less as far were the highest among three kinds of plans. V5 , V10 of the infected
as possible. The three radiotherapy plans are as follows: lung were slightly higher in VMAT and IMRT. Each evaluation index
of the contralateral lung and heart in 3D-CRT were the lowest. D1
(a) 3D-CRT: using 6 MV X-ray, chest wall using 1/4 tangential field, of the contralateral breast were lower in IMRT and VMAT plans,
lymphatic drainage area around collar bone using isocenter which were 1770.89 ± 121.16 cGy and 1839.92 ± 92.77 cGy, respec-
semi field irradiation. Avoid shoot omitting and dosage overlap- tively. While D1% for the spinal cord in IMRT and VMAT plans were
ping. And algorithm model is Anisotropic Analytical Algorithm higher, which were 1990.12 ± 61.52 cGy and 1927.38 ± 43.67 cGy,
(AAA). respectively.
(b) IMRT: using 6 MV X-ray, 5 fields isocenter way (130◦ , 95◦ , 0◦ ,
330◦ , 290◦ ) to do the reversal dynamic optimization design. And 3.3. Normal tissues comparison
algorithm model is AAA.
(c) VMAT: using 6 MV X-ray, double arc way (clockwise and coun- As shown in Fig. 2, in the range of 500–1500 cGy, the normal
terclockwise) to disperse field, abduction of 10◦ −25◦ by tangent tissue volume was lowest in 3D-CRT plan, while the volume was
H. Liu et al. / Computerized Medical Imaging and Graphics 54 (2016) 1–5 3

Table 1
Dosimetric Parameters Comparison of CTV and OARs in three plans (3D-CRT, IMRT, VMAT).

Project 3D-CRT IMRT VMAT P value

CTV
HI 1.12 ± 0.11 1.10 ± 0.09 1.10 ± 0.10 –
CI 0.33 ± 0.08 0.77 ± 0.05 0.79 ± 0.05 a,b
Dmean (cGy) 4944.03 ± 92.30 5139.32 ± 56.32 5051.09 ± 61.91 a,b,c
D2% (cGy) 5299.38 ± 51.82 5455.51 ± 39.11 5358.88 ± 44.63 –
D98 (cGy) 4518.21 ± 103.40 4753.62 ± 66.23 4768.93 ± 63.47 a,b
V90 (%) 98.24 ± 0.51 99.28 ± 0.72 99.29 ± 0.37 a,b
V95 (%) 90.22 ± 1.20 96.05 ± 0.77 97.44 ± 1.21 a,b,c

Infected lung
Dmean (cGy) 1385.14 ± 110.34 1353.21 ± 98.43 1378.37 ± 83.42 –
V5 (%) 41.67 ± 4.20 72.36 ± 5.71 76.53 ± 4.29 a,b,c
V10 (%) 35.58 ± 3.37 41.99 ± 3.01 45.02 ± 2.74 a,b,c
V20 (%) 29.45 ± 2.46 24.46 ± 2.05 25.41 ± 1.98 a,b
V30 (%) 28.14 ± 1.34 15.95 ± 1.27 16.17 ± 0.88 a,b

Contralateral lung
Dmean (cGy) 107.77 ± 18.23 466.22 ± 34.45 532.18 ± 42.31 a,b,c
V5 (%) 2.08 ± 0.31 35.89 ± 2.55 37.47 ± 2.45 a,b
V10 (%) 0.92 ± 0.14 11.85 ± 1.34 10.58 ± 1.57 a,b

Contralateral breast
Dmean (cGy) 624.71 ± 50.96 667.57 ± 64.81 844.65 ± 57.29 a,b,c
D1 (cGy) 3934.77 ± 301.47 1770.89 ± 121.16 1839.92 ± 92.77 a,b

heart
Dmean (cGy) 881.53 ± 51.56 1302.42 ± 67.22 1367.39 ± 61.64 a,b
V10 (%) 20.11 ± 1.45 46.94 ± 2.06 45.27 ± 2.31 a,b
V5 (%) 15.58 ± 1.82 17.55 ± 1.49 18.94 ± 1.23 a,b

Spinal cord
D1 (cGy) 807.92 ± 145.57 1990.12 ± 61.52 1927.38 ± 43.67 a,b

Notes: a = 3D-CRT vs. IMRT; b = 3D-CRT vs. VMAT; c = IMRT vs. VMAT.

Fig. 1. The three dose distribution of the cases based on one patient sample.

highest in VMAT plans; the volumes in three plans were similar plan was 94 s less than that in IMRT plan, which remarkably having
in V20 , and the volumes in IMRT and VMAT plans become similar improved the utilization of X-rays.
starting from 1500 cGy; compared with the other two plans, the
volume in 3D-CRT plan was highest at the point of 3000 cGy. 4. Discussion

VMAT is a kind of Rotational Intensity Modulated technology


3.4. MU and treatment time comparison based on VMAT theory proposed by Otto (Otto, 2008). The full
arc frame can rotate about 360◦ . The arc consists of 177 control
The average MU were 496 ± 27, 827 ± 31 and 408 ± 16, respec- nodes, whereby the rotating speed of the frame is 4.8◦ /s. The max-
tively in 3D-CRT, IMRT and VMAT plans. MU in VMAT plan was imum dose rate is 600 MU/min. MLC blade’s maximum speed is
50.67% of that in IMRT. Moreover, the treatment time were 172s, 2.5 cm/s. Gantry rotation takes about 75 s per circle. Many scholars
213s, and 119s, respectively. However, the treatment time in VMAT have done researches on VMAT in the body, the head and the neck.
4 H. Liu et al. / Computerized Medical Imaging and Graphics 54 (2016) 1–5

scattering radiation is more and field passes through the normal


tissue.
In general, the results shows that the ipsilateral lung V20 of the
three plans is not significantly different. VMAT plan can signifi-
cantly reduce the high dose volume of the ipsilateral lung (V30 ).
But due to the increasing of scattered radiation, IMRT and VMAT
plan also significantly increase the low dose irradiated volume (V5 ,
V10 ) compared to 3D-CRT plan. VMAT plan has more average dose
to the contralateral lung, contralateral breast and heart than 3D-
CRT plans. But we still cannot definitely tell the advantages and
disadvantages of the three plans on lung protection. The is because
the DVH parameters can be related to the radiation injury based
on different angles. According to DVH data of the radiotherapy for
lung cancer patients, the risk of radiation pneumonitis is related
to the average dose of lung (MLD) and V20 , V30 (Changsheng and
Fig. 2. The histogram of low dose area of the normal tissue in three radiotherapy
plans.
Liu Tonghai, 2010; Hernando et al., 2001; Kwa et al., 1998). And
V5 , V10 are to the effective factors to estimate the occurrence of
radiation pneumonitis (Heyi and Xu Bingqing, 2009). This paper
The results show that VMAT can reduce the total time of radiother- demonstrates VMAT plan protects the normal tissue of the affected
apy plan for patients and the beam-on time of the accelerator. The side with good effect.
greatest advantage of VMAT technology is to further reduce the The results are also similar to the results by Qiu et al. (2010),
treatment time and the number of MU without reducing dose dis- Shaitelman et al. (2011), and Sun Tao and Min (2012). The three
tribution, so as to improve the treatment target of biological effects plans have not much difference for the heart. But for the spinal cord,
and the number of patients treated in a unit of time (Li et al., 2009; contralateral lung and contralateral breast, VMAT has no advan-
Jagsi et al., 2010; Shaffer et al., 2009; Jinhu and Liu Tonghai, 2010). tage. VMAT plan has minimum number of monitor units MU and
Because the number of MU reduces obviously, thereby reducing the shortest treatment time. For left breast cancer patients, 3D-CRT
number of scattering lines of the accelerator head collimator, the can meet the clinical requirements, but has no advantage in the
risk of cancer reoccurrence is reduced theoretically. protection of normal breast tissue. VMAT plan has the advantages
The study compares three kinds of radiotherapy techniques for of protecting normal breast tissue, and can obviously shorten the
left breast cancer. These three treatments can meet the clinical treatment time. And the influence of the movement during treat-
requirements. The target area fitness and DVH of 3D-CRT were not ment, organ changes caused by respiratory motion and involuntary
so good as those of IMRT and VMAT, which may be related to the movement will be reduced accordingly. Patient’s discomfort and
field conditions of 3D-CRT. But the doses delivered to the spinal graded internal displacement are reduced. Ultimately the accuracy
cord D1 and the infected lung (including V5 , V30 ) in 3D-CRT were of dose distribution and treatment effect are improved. Meanwhile,
the lowest. But the average dose Dmean , V20 and V30 were higher in the process of designing and optimizing IMRT and VMAT Plans, it
compared with the other two plans. Furthermore, the average doses takes longer time than that of 3D-CRT because the parameters are
to contralateral lung, contralateral breast and heart were the lowest adjusted and optimized repeatedly. Especially in the VMAT plan,
while the high dose point was higher. Because such field mode can the optimization process is divided into 5 steps, one by one com-
maximally avoid the spinal cord, but lead to increasing the high pleted. It not only needs to optimize the sub field and weight, but
dose area and the average dose of ipsilateral lung. IMRT also has also because of many physical parameters in optimization plan, the
good dose distribution in the target. It can reduce the maximum optimization process is complex limited to the version of the sys-
dose of the target area, and allow the average dose Dmean of tar- tem. The time-consuming process can reduce our work efficiency
get area to be closer to the prescribed dose. The target area fitness considerably.
and DVH can meet the clinical needs. But its reception amount of
the spinal cord D1 is higher than that in 3D-CRT and VMAT plans
and the reception amount of ipsilateral lung (including V20 , V30 , the 5. Conclusion
average dose Dmean ) is the lowest. But the V5 and V10 are between
3D-CRT and VMAT plans. Also, the average doses to the contralat- In our paper, we note that all of the three plans based on 3D-CRT,
eral lung, contralateral breast and the heart are between 3D-CRT IMRT and VMAT technology can achieve the basic requirements of
and VMAT plans. The target area fitness and DVH of VMAT are bet- clinical treatment, but in the process of treatment there are many
ter than the other two plans, and the reception amount of the spinal uncertain factors. Therefore we need to obtain an accurate target
cord D1 is between the other two plans. All indicators of ipsilateral volume delineation and strict control in order to assure a high qual-
lung were higher than that of other plans, and some other indicators ity. The dosimetric parameters pertaining to the Rapid Arc and IMRT
(Dmean , V5 ) of contralateral lung, contralateral breast and heart are have certain advantages. In addition, they have greatly increased
higher than the other plans. In addition, the indicator (V10 ) is better the number of MU, the efficiency of treatment and time. But they
than the other plans. This study showed that IMRT and VMAT has need to further reduce the amount of the subject in the contralateral
incomparable advantages than 3D-CRT plan in dose distribution lung, heart, spinal cord, and the other organs.
and uniformity. They can guarantee the treatment target to obtain
sufficient dose, reducing the cold and hot points of the dose in the Conflict of interest
target area. That can prevent tumor of chest wall recurrence. For
the high dose region volume of the normal tissue, IMRT and VMAT The authors declared no conflict of interest.
plans are smaller, while 3D-CRT is larger; For the low dose region
volume of normal tissue, IMRT and VMAT plans are larger, and 3D-
CRT plan is smaller. The reason of having larger low dose volume Role of the funding source
of normal tissues in IMRT and VMAT plans may be the following:
the number of intensity modulated radiation field is more, field This research is not funded.
H. Liu et al. / Computerized Medical Imaging and Graphics 54 (2016) 1–5 5

References Otto, K., 2008. Volumetric modulated arc therapy: IMRT in a single gantry arc. Med.
Phys. 35 (1), 310–317.
Anon, 2010. Prescribing, recording, and reporting photon-beam Palma, D., Vollans, E., James, K., et al., 2008. Volumetric modulated arc therapy for
intensity-modulated radiation therapy (IMRT): contents. J. ICRU 10 (1), 1–106. delivery of prostate radiotherapy: comparison with intensity-modulated
Changsheng, Ma, Liu Tonghai, Y.Y., 2010. Dosimetric comparison between rapid radiotherapy and three-dimensional conformal radiotherapy. Int. J. Radiat.
arc and fixed gantry intensity modulated radiation of liver carcinoma. Chin. J. Oncol. Biol. Phys. 72 (4), 996–1001.
Radiat. Med. Prot. 30 (5), 581–584. Qiu, J.J., Chang, Z., Wu, Q.J., et al., 2010. Impact of volumetric modulated arc
Hernando, M.L., Marks, L.B., Bentel, G.C., et al., 2001. Radiation-induced pulmonary therapy technique on treatment with partial breast irradiation. Int. J. Radiat.
toxicity: a dose-volume histogram analysis in 201 patients with lung cancer. Oncol. Biol. Phys. 78 (1), 288–296.
Int. J. Radiat. Oncol. Biol. Phys. 51 (3), 650–659. Shaffer, R., Nichol, A.M., Vollans, E., 2009. A comparison of volumetric modulated
Heyi, Fu, Xu Bingqing, L.B., 2009. Prospective clinical study of V5 and V10 in arc therapy and conventional intensity-modulated radiotherapy for frontal
predicting radiation-induced lung injury and three dimensional conformal and temporal high-grade gliomas. Int. J. Radiat. Oncol. Biol. Phys. 76 (4),
radiation therapy for non small cell lung cancer in stage III and IV. Chin. J. 1177–1184.
Radiat. Oncol. 18 (6), 439–442. Shaitelman, S.F., Kim, L.H., Yan, D., et al., 2011. Continuous arc rotation of the couch
Jagsi, R., Jm, B.D.M., Marsh, R.B., et al., 2010. Unacceptable cosmesis in a protocol therapy for the delivery of accelerated partial breast irradiation: a treatment
investigating intensity-modulated radiotherapy with active Breathing control planning analysis. Int. J. Radiat. Oncol. Biol. Phys. 80 (3), 771–778.
for accelerated partial-breast irradiation. Int. J. Radiat. Oncol. Biol. Phys. 76 (1), Sun Tao, L.J., Xu, Min, 2012. Dosimetric comparison of 3D-CRT, dIMRT, and
71–78. RapidArc technology in partial external irradiation of breast. Chin. J. Radiat.
Jinhu, Chen, Liu Tonghai, Y.Y., 2010. Comparative study of radiotherapy planning Med. Prot. 32 (1), 74–79.
for cervical esophageal cancer patients between fixed gantry intensity Vicini, F.A., Yan, D., Matter, R.C., 2000. Intensity modulation to improve dose
modulated radiation and RapidArc. Chin. J. Radiat. Oncol. 19 (5), 429–433. uniformity with tangential breast radiotherapy: initial clinical experience. Int.
Kwa, S.L.S., Lebesque, J.V., Theuws, J.C.M., et al., 1998. Radiation pneumonitis as a J. Radiat. Oncol. Biol. Phys. 48 (5), 1559–1568 (10).
function of mean lung dose: an analysis of pooled data of 540 patients. Int. J. Yin, Y., Chen, J., Xing, L., et al., 2011. Applications of IMAT in cervical esophageal
Radiat. Oncologybiol. 42 (1), 1–9. cancer radiotherapy: a comparison with fixed-field IMRT in dosimetry and
Li, J.B., Wang, J.G., Lu, J., et al., 2009. Influence of active breathing control on the implementation. J. Appl. Clin. Med. Phys. 12 (12), 48–57.
dose distribution in the target of forward whole-breast intensity-modulated
radiotherapy after breast conserving surgery. Zhonghua zhong liu za zhi [Chin.
J. Oncol.] 31 (8), 617–621.