Application of carbon nanoparticles in lymph node dissection and parathyroid protection during thyroid cancer surgeries: a systematic review and meta-analysis

Abstract

Purpose

To investigate whether carbon nanoparticles (CNs) are helpful in identifying lymph nodes and metastatic lymph nodes and in parathyroid protection during thyroid cancer surgery.

Methods

English and Chinese literature in PubMed, Cochrane Database of Systematic Reviews, EMBASE, ClinicalTrials.gov, China Biology Medicine Database, China National Knowledge Infrastructure, China Master’s and Doctoral Theses Full-Text Database, Wanfang database, and Cqvip database were searched (till March 22, 2016). Randomized controlled trials (RCTs) that compared the use of CNs with a blank control in patients undergoing thyroid cancer surgery were included. Quality assessment and data extraction were performed, and a meta-analysis was conducted using RevMan 5.1 software. The primary outcomes were the number of retrieved central lymph nodes and metastatic lymph nodes, and the rate of accidental parathyroid removal.

Results

We obtained 149 relevant studies, and only 47 RCTs with 4,605 patients (CN group: n=2,197; blank control group: n=2,408) met the inclusion criteria. Compared with the control group, the CN group was associated with more retrieved lymph nodes/patient (weighted mean difference [WMD]: 3.39, 95% confidence interval [CI]: 2.73–4.05), more retrieved metastatic lymph nodes (WMD: 0.98, 95% CI: 0.61–1.35), lower rate of accidental parathyroid removal, and lower rates of hypoparathyroidism and hypocalcemia. However, the total metastatic rate of the retrieved lymph nodes did not differ between the groups (odds ratio: 1.13, 95% CI: 0.87–1.47, P=0.35).

Conclusion

CNs can improve the extent of neck dissection and protect the parathyroid glands during thyroid cancer surgery. And the number of identified metastatic lymph nodes can be simultaneously increased.

Keywords:

• carbon nanoparticles
• lymph node tracer
• thyroid cancer
• parathyroid
• meta-analysis

Introduction

Thyroid cancer is a common malignant tumor, and its occurrence has been increasing appreciably over the last few decades.^1,2 Papillary thyroid carcinoma, the most common pathological type of thyroid carcinoma, is associated with an excellent prognosis if surgery is performed successfully.^3,4 However, complications and tumor metastasis often occur postoperatively. A multicenter revisit study of 25,634 patients with a history of ambulatory surgery revealed that the incidence of hypocalcemia was as high as 20.8%.⁵ Moreover, the rate of postoperative cervical lymph node metastasis has been reported to be as high as 20%–90%.^6,7 Therefore, an effective method is urgently required to help identify and remove additional lymph nodes and decrease the risk of parathyroid injury.

Nanobiotechnology, a new field defined as biomedical applications of nanosized systems, which involves nanostructure and nanomaterials, has emerged as a key player among various disciplines of biomedical science. Nanomaterials, which measure 1–1,000 nm, have unique physical and chemical properties such as small-size effect, large surface area, high reactivity, and quantum effects. In addition, they have been certified as breaking a new ground in disease detection,^8,9 imaging,^10,11 diagnosis,^12,13 and treatment.^12,14–16 Carbon-based nanoparticles are an important part of nanomaterials; they include carbon nanotubes, fullerene, and graphene and its derivatives.¹⁷ Due to their unique physical and chemical properties, carbon-based nanoparticles have broad applications in the biomedical field. Above all, structure or surface modifications of carbon-based nanoparticles result in different effects.

In recent years, with the development of nanotechnology, nanomaterials or nanosized products have been used in surgeries. Carbon nanoparticles (CNs) suspension (China Food and Drug Administration approval H20041829; Lai Mei Pharmaceutical Co, Chongqing, People’s Republic of China), which comprises nanosized polymeric carbon granules with an average diameter of 150 nm (Figure 1),¹⁸ ensures that these CNs pass through the lymphatic vessels (diameter: 120–500 nm) rather than blood capillaries (diameter: 20–50 nm) due to their molecular size. Hence, the lymph nodes and thyroid glands can be stained with CNs but not the parathyroid glands. Therefore, as revealed in Figure 2,¹⁹ CNs have been used as lymph node tracers during thyroid surgeries in the People’s Republic of China in recent years, but there is no report about their usage in other countries. Previous studies have demonstrated that CNs could help visualize the lymph nodes and preserve the parathyroid. In addition, a meta-analysis has also been published to support this viewpoint.²⁰ However, controversy still exists about the usefulness of CNs. Liu et al reported that CNs are not beneficial for parathyroid protection during thyroid cancer surgery and that the usage of CNs results in a significantly prolonged operation time.¹⁸ Before 2014, all articles published on the usage of CNs in thyroid cancer surgeries were in Chinese, and there were only limited randomized controlled trials (RCTs) reported. Accordingly, a considerable amount of non-RCTs were included in the previous meta-analysis.²⁰ However, in recent years, more RCTs have been published. Hence, another meta-analysis is needed to determine whether CNs are helpful in thyroidectomy.

Figure 1 Transmission electron microscopic image of CNs.

Notes: The main active component of CNs is nanosized polymeric carbon granules with an average diameter of 150 nm. Republished with permission of SAGE Publications, Inc., from Liu X, Chang S, Jiang X, Huang P, Yuan Z. Identifying parathyroid glands with carbon nanoparticle suspension does not help protect parathyroid function in thyroid surgery: a prospective, randomized control clinical study. Surg Innov. 2016;23(4):381–389. © The Author(s) 2016; permission conveyed through Copyright Clearance Center, Inc.¹⁸
Abbreviation: CN, carbon nanoparticle.

Figure 2 Intraoperative view of the black-stained clusters of lymph nodes by CN injection.

Notes: Reproduced with permission of John Wiley and Sons, from Zhu Y, Chen X, Zhang H, et al. Carbon nanoparticle-guided central lymph node dissection in clinically node-negative patients with papillary thyroid carcinoma. Head Neck. 2016;38(6):840–845. © 2015 Wiley Periodicals, Inc.¹⁹
Abbreviation: CN, carbon nanoparticle.

With this in mind, the present systematic review and meta-analysis was designed to confirm whether CNs are indeed helpful in thyroid cancer surgery, that is, whether CNs can really improve the extent of thyroidectomy and neck dissection and help identify metastatic lymph nodes while preserving the parathyroid glands, as compared with the performance of blank controls.

Methods

Search strategy

The following English and Chinese databases were searched systematically by 2 investigators independently (till March 22, 2016): PubMed, Cochrane Database of Systematic Reviews, EMBASE, ClinicalTrials.gov, China Biology Medicine Database, China National Knowledge Infrastructure, China Master’s and Doctoral Theses Full-Text Database, WANFANG database, and Cqvip database. RCTs on initial thyroid cancer surgeries that compared the use of CNs with a blank control were included. Our search terms included (nano-carbon) or (carbon particle) or (carbon nanoparticle) or (carbon nanoparticles) or (lymph node tracer) or (lymphatic tracer) or (lymphography) and (thyroid or thyroidea). To resolve any disagreement between the 2 investigators, a third reviewer was invited to assess any discrepant items.

Inclusion criteria

The studies selected were RCTs on thyroid cancer surgeries that included: 1) patients who underwent initial surgery and with a confirmed pathology diagnosis and 2) a control group not injected with anything before thyroidectomy and a CN group injected with CNs.

Exclusion criteria

Non-independent clinical controlled trials, non-RCTs, studies with a patient number <10, or studies with incomplete data were excluded.

Observation indexes

The primary outcomes were the number of retrieved central lymph nodes and metastatic lymph nodes per patient, and the rate of accidental parathyroid removal. Other outcomes extracted from the identified RCTs included the staining rate of lymph nodes, the number of metastatic lymph nodes in all retrieved lymph nodes, and the rate of postoperative transient or permanent hypoparathyroidism and hypocalcemia. All included articles reported at least 1 of the outcomes.

Quality assessment

To evaluate the quality of these studies, the Jadad scoring system was applied in the RCTs. The scoring system included 3 items: descriptions of the dropouts and withdrawals (0 or 1 point), blinding (0–2 points), and randomization (0–2 points). The maximum score was 5 points. RCTs that scored 3–5 points were considered to be of high quality, whereas a score of 0–2 was considered to indicate low quality.

Statistical analysis

RevMan version 5.1 software was used for the statistical analyses. We measured the heterogeneity of the studies using the I² and χ² tests. Statistical heterogeneity of the studies was defined as an I² value <50% or P-value <0.10. The random-effects model was applied if heterogeneity existed among the studies; otherwise, the fixed-effects model was adopted for the analyses. Weighted mean differences (WMDs) with 95% confidence intervals (CIs) were used for the continuous outcome variables, whereas risk differences (RDs) and odds ratios (ORs) with 95% CIs were calculated for the dichotomous outcome variables. To investigate possible bias, funnel plots were created. For all analyses, statistical differences were considered to exist between the 2 groups when P-value was <0.05.

Results and discussion

Considering the complicated anatomic structure and lymphatic drainage, the cervical lymph node metastasis rate is amazingly high after thyroid cancer surgery.^6,7 Especially, in the central region of the neck and other places such as the central neck compartment and deep surface of the recurrent laryngeal nerve, which cannot be easily dissected, postoperative cervical lymph node metastasis is common. As a result, reoperation and surgical trauma are common in these patients. According to a previous report,²¹ while the incidence of permanent hypoparathyroidism is 3%–10% after the first surgery for thyroid disease, it is as high as 9%–35% after reoperation. Kurmann et al reported that the incidence of permanent recurrent laryngeal nerve palsy was significantly higher in patients undergoing reoperation on the ipsilateral lobe compared to patients undergoing initial operation (3.8% vs 1.1%; P=0.03).²² In addition, parathyroid injury has been considered inevitable for a long time, mostly due to its unique anatomy. The appearance of the parathyroid is similar to that of the cervical lymph nodes, and the location of the gland is close to the backside of the thyroid gland and varies greatly; for example, the parathyroid may hide within the thyroid lobes, thymus, or carotid sheath. Xu and Gu reported that 6.9%–46% of parathyroid glands were damaged during thyroid surgery.²³ Such damage may cause permanent or transient hypocalcemia and hypoparathyroidism, and will consequently affect the quality of life of the patients. Therefore, some technical methods are urgently needed to help us better visualize the lymph nodes and metastatic lymph nodes, and distinguish the parathyroid glands and preserve them.

CNs, which have a mean diameter of 150 nm and a lymphatic tendency, had been used as lymph node tracers clinically in other cancer surgeries before. Upon injection, CNs are rapidly devoured by macrophages, resulting in the lymph nodes, but not the capillaries, initially becoming black-stained. Subsequently, the thyroid tissue stains black, as do the surrounding lymph nodes, whereas any tissues without lymph vessel connections remain unstained. Some previous studies have demonstrated that CNs are beneficial for visualizing the lymph nodes and for distinguishing and preserving the parathyroid glands. On the contrary, other studies found no advantage of CNs.^18,24 In addition, some indexes, such as postoperative hypocalcemia, are easily influenced by confounding factors such as the postoperative therapeutic selection, calcium supplements, and individual differences. To date, no large-scale meta-analysis has been performed to clarify these diverging results. Therefore, our analysis was designed to resolve the problems.

Literature search and study description

According to the search strategy, 149 references were obtained, and 47 RCTs met the inclusion criteria.^{18,19,23–67} The flowchart of the literature search is shown in Figure 3. A total of 4,605 patients were included in the analysis, including 2,197 patients in the CN group and 2,408 patients in the blank control group. All patients had confirmed thyroid cancer by postoperative pathologic diagnosis and had been divided into the CN and blank control groups randomly. All of the CNs used in these studies were the same product.

Figure 3 Flowchart of the literature search.

The characteristics of all 47 RCTs are presented in Table 1. The Jadad scale system was used to assess the quality of the included studies (Table 2). Most investigators preferred multipoint injections (2–4 points) prior to the thyroidectomy, with a total dose of approximately 0.3–0.8 mL. Besides, according to our experience and the studies analyzed herein, no adverse reactions to CNs have been reported.

Table 1 Characteristics of the 47 RCTs included in the meta-analysis

Study	n	Male/female		Age, years, mean (SD)		Injection site	Dose (mL)	Waiting time	Staining rate (%)	Indicesa
		CNs	BL	CNs	BL
Chen et al (2016)^Citation25	173	15/72	13/73	33.46 (4.58)	34.32 (4.36)	Upper, middle, and lower points of the thyroid	0.3	5–10 min	NA	1, 4–6
Zhang et al (2016)^Citation26	37	3/14	3/17	41.1 (8.7)	43 (7.2)	In the bilateral thyroid	0.2–0.4	5–20 min	95.5	1–4
Liu et al (2016)^Citation18	156	16/62	17/61	NA	NA	Upper, middle, and lower parts of the lobes	0.4–0.8	NA	NA	4, 6
Li (2015)^Citation27	40	4/16	8/12	47.9 (11.2)	49.5 (12.4)	Top of the tumor	0.4–0.8	3 min	NA	4
Feng and He (2015)^Citation28	60	0/30	0/30	NA	NA	4 points around the tumor	0.8	NA	NA	4–6
Guo et al (2015)^Citation29	59	9/21	8/21	35 (9.5)	32 (8.2)	Upper, middle, and lower points of the bilateral thyroid	0.6	20 min	NA	6
Duan et al (2015)^Citation30	80	14/26	15/25	40.4 (7.18)	42.5 (7.65)	3–5 points around the thyroid	0.3–0.5	30 min	95.7	1, 2, 4, 5
Chen and Wu (2015)^Citation31	96	11/37	16/32	46.1 (10.9)	45.8 (12.6)	NA	NA	NA	NA	1, 5, 6
Du et al (2015)^Citation32	118	19/41	18/40	42.7 (10.6)	43.5 (10.8)	Upper and middle points of the thyroid	0.4	5 min	NA	4–6
Wang et al (2015)^Citation33	120	NA	NA	NA	NA	Upper and middle points of the thyroid	0.2	15 min	NA	1–6
Liu and Qing (2015)^Citation34	66	12/21	11/22	46.11 (2.09)	45.63 (2.7)	Upper, middle, and lower points of the thyroid	0.3	20 min	NA	4–6
Shao et al (2015)^Citation35	60	NA	NA	NA	NA	Multipoint injection of the tumor	0.4–1.2	10 min	NA	1, 5, 6
Fu (2015)^Citation43	250	NA	NA	NA	NA	4–6 points around the tumor	0.8–1.2	15–20 min	NA	5
Wu et al (2015)^Citation36	245	11/79	33/122	NA	NA	Top and middle points of the tumor	0.2–0.6	5–10 min	NA	4–6
Li et al (2015)^Citation37	52	NA	NA	NA	NA	4 points of the bilateral thyroid	1.0	10 min	NA	4–6
Chu et al (2015)^Citation38	57	10/18	8/21	46.28 (12.635)	40.39 (1.711)	Upper, middle, and lower points of the ipsilateral thyroid	0.3–0.6	10 min	NA	1, 4
Wang et al (2015)^Citation39	88	10/34	9/35	36.6 (11.2)	36.8 (11.4)	Upper, middle, and lower points of the thyroid	0.3–0.8	NA	NA	1
Wu et al (2015)^Citation40	86	NA	NA	NA	NA	Upper, middle, and lower points of the thyroid	0.6–1.2	30 min	82.9	1–3
Yin et al (2015)^Citation41	80	17/23	18/22	NA	NA	Upper, middle, and lower points of the thyroid	0.6	5 min	NA	4–6
Li et al (2015)^Citation42	72	9/27	10/26	NA	NA	2–4 points around the thyroid	NA	15 min	84	1–6
Xu and Gu (2016)^Citation23	114	5/52	4/53	45.37 (10.71)	42.68 (14.43)	Around the tumor	0.5	5–10 min	NA	1, 4–6
Wang et al (2015)^Citation64	55	1/27	2/25	30.25 (6.04)	29.44 (6.27)	In the thyroid gland	0.1–0.2	2–3 min	85	1, 4–6
Zhu et al (2016)^Citation19	162	14/67	16/65	46.75 (12.09)	44.31 (10.73)	1–2 points in the thyroid gland	0.1–0.2	Few minutes	92.75	1–6
Gu et al (2015)^Citation65	100	10/40	6/44	46.98 (9.027)	47.76 (13.912)	Upper and lower points of the thyroid	0.2–0.3	3–5 min	NA	1, 2, 4–6
Liu et al (2014)^Citation44	47	3/20	5/19	37.79 (11.2)	33.94 (7.76)	Upper, middle, and lower points of the thyroid	0.6	1 day	NA	1, 4–6
Chen et al (2014)^Citation45	72	5/31	8/28	38.23 (10.67)	34.64 (8.75)	Upper, middle, and lower points of the thyroid	0.4/0.6	30 min	95.4	1, 2, 4
Gao and Zhao (2014)^Citation46	100	12/38	9/41	42 (1.28)	41 (1.36)	3–4 points around the tumor	0.53–1	30 min	95.56	1–5
Yang et al (2014)^Citation47	379	23/155	27/174	42.3 (8.5)	44.2 (7.3)	Upper and lower points of the tumor or the thyroid gland	0.1–0.3	3–5 min	NA	1, 6
Wang and Rang (2014)^Citation48	70	NA	NA	44.3 (8.8)	45.2 (7.9)	4 points in the contralateral thyroid gland	0.4–0.8	15–20 min	NA	1, 2, 4–6
Liu et al (2014)^Citation49	55	5/21	9/20	36.58 (11.31)	33.58 (7.77)	In the ipsilateral thyroid	14 days	NA	NA	1, 4–6
Du (2014)^Citation50	40	4/16	8/12	47.9 (10.12)	49.5 (12.4)	Top point of the thyroid gland	0.4–0.8	3 min	NA	1, 4–6
Zhao (2014)^Citation51	183	23/79	19/62	NA	NA	Upper and lower points of the thyroid	0.1–0.2	5 min	73.5	1, 2
Chun (2014)^Citation52	67	6/27	11/23	41.06 (12.84)	44.65 (12.84)	Upper, middle, and lower points of the thyroid	0.1–0.3	3–5 min	86.8	1, 5
Liu (2014)^Citation53	184	23/53	30/78	46.5 (12.8)	47.2 (13.5)	Multipoint injection of the tumor	0.8	NA	NA	5, 6
Zhang et al (2014)^Citation54	72	12/24	10/26	NA	NA	Upper and lower points of the thyroid	0.2–0.4	5 min	NA	4–6
Shen et al (2014)^Citation55	109	NA	NA	NA	NA	Around the tumor	0.2	10 min	90.5	1, 2, 4
Shao et al (2014)^Citation57	29	NA	NA	NA	NA	Upper, middle, and lower points of the thyroid	0.3–0.6	15 min	NA	4–6
Long et al (2014)^Citation56	150	15/60	12/63	41.4 (1.62)	42.1 (2.56)	Upper, middle, and lower points of the ipsilateral thyroid	0.3	20 min	NA	1, 2, 6
Tian et al (2014)^Citation66	100	5/45	11/39	36.4 (2.5)	44.5 (5.8)	Upper, middle, and lower points of the thyroid	0.6	10–15 min	NA	1, 2, 4–6
Sun et al (2014)^Citation67	80	NA	NA	NA	NA	2–4 points around the tumor	0.2–0.8	15 min	69.89	1–4
Yang et al (2013)^Citation58	43	1/20	1/21	32.48 (4.69)	32.32 (5.35)	In the ipsilateral thyroid gland	0.1	20 min	NA	1, 2, 4, 5
Yang et al (2013)^Citation59	68	17/19	12/20	34.5 (9.1)	33.9 (10.3)	1–2 points around the tumor	0.1–0.3	10 min	91.7	1, 5
Wu (2013)^Citation60	55	5/21	9/20	36.58 (11.31)	33.58 (7.77)	Upper, middle, and lower points of the ipsilateral thyroid	0.6	NA	NA	1, 2, 4–6
Huang et al (2013)^Citation24	72	12/24	10/26	41.22 (2.53)	40.69 (2.42)	Lower and upper points of the tumor	0.2–0.4	NA	NA	5
Bai et al (2013)^Citation61	88	9/39	7/33	46.28 (9.2)	45.39 (12.03)	Lower and upper points of the tumor	0.2	NA	NA	1–6
Zeng et al (2012)^Citation62	80	0/40	0/40	NA	NA	4–6 points around the tumor	1.0	20 min	NA	5, 6
Wang et al (2009)^Citation63	36	10/8	7/11	NA	NA	4–6 points around the tumor	0.4–0.6	30 min	95.2	2–4

Notes:

a Observation indexes: differences in the following: 1) the number of retrieved lymph nodes, 2) the total metastatic rate of the retrieved lymph nodes, 3) the metastatic rate of stained/unstained lymph nodes, 4) the accidental parathyroid removal rate, 5) the postoperative transient or permanent hypoparathyroidism rate, and 6) the postoperative transient or permanent hypocalcemia rate between the CN and blank control groups.

Abbreviations: RCT, randomized controlled trial; SD, standard deviation; CN, carbon nanoparticle; BL, blank; NA, not available.

Table 2 Quality assessment of the 47 randomized controlled trials included using the Jadad scale system

Study	Randomization	Concealment of allocation	Blinding	Loss to follow-up (%)	Quality assessment
Chen et al (2016)^Citation25	No detailed description	Only mentioned randomized	Unclear	0	2
Zhang et al (2016)^Citation26	No detailed description	Only mentioned randomized	Unclear	0	2
Liu et al (2016)^Citation18	Computer-generated permuted	Only mentioned randomized	Unclear	0	3
Li (2015)^Citation27	No detailed description	Only mentioned randomized	Unclear	0	2
Feng and He (2015)^Citation28	No detailed description	Only mentioned randomized	Unclear	0	2
Guo et al (2015)^Citation29	No detailed description	Only mentioned randomized	Unclear	0	2
Duan et al (2015)^Citation30	No detailed description	Only mentioned randomized	Unclear	0	2
Chen and Wu (2015)^Citation31	No detailed description	Only mentioned randomized	Unclear	0	2
Du et al (2015)^Citation32	No detailed description	Only mentioned randomized	Unclear	0	2
Wang et al (2015)^Citation33	No detailed description	Only mentioned randomized	Unclear	0	2
Liu and Qing (2015)^Citation34	No detailed description	Only mentioned randomized	Unclear	0	2
Shao et al (2015)^Citation35	No detailed description	Only mentioned randomized	Unclear	0	2
Fu (2015)^Citation43	No detailed description	Only mentioned randomized	Unclear	0	2
Wu et al (2015)^Citation36	No detailed description	Only mentioned randomized	Unclear	0	2
Li et al (2015)^Citation37	Random-number table	Only mentioned randomized	Unclear	0	3
Chu et al (2015)^Citation38	No detailed description	Only mentioned randomized	Unclear	0	2
Wang et al (2015)^Citation39	No detailed description	Only mentioned randomized	Unclear	0	2
Wu et al (2015)^Citation40	Random-number table	Only mentioned randomized	Unclear	0	3
Yin et al (2015)^Citation41	Random-number table	Only mentioned randomized	Unclear	0	3
Li et al (2015)^Citation42	Random-number table	Only mentioned randomized	Unclear	0	3
Xu and Gu (2016)^Citation23	No detailed description	Only mentioned randomized	Unclear	0	2
Wang et al (2015)^Citation64	Computer-generated	Only mentioned randomized	Unclear	0	3
Zhu et al (2016)^Citation19	Computer-generated random-number tables	Only mentioned randomized	Unclear	0	3
Gu et al (2015)^Citation65	No detailed description	Only mentioned randomized	Unclear	0	2
Liu et al (2014)^Citation44	No detailed description	Only mentioned randomized	Unclear	0	2
Chen et al (2014)^Citation45	No detailed description	Only mentioned randomized	Unclear	0	2
Gao and Zhao (2014)^Citation46	No detailed description	Only mentioned randomized	Unclear	0	2
Yang et al (2014)^Citation47	No detailed description	Only mentioned randomized	Unclear	0	2
Wang and Rang (2014)^Citation48	Random-number table	Only mentioned randomized	Unclear	0	3
Liu et al (2014)^Citation49	No detailed description	Only mentioned randomized	Unclear	0	2
Du (2014)^Citation50	No detailed description	Only mentioned randomized	Unclear	0	2
Zhao (2014)^Citation51	No detailed description	Only mentioned randomized	Unclear	0	2
Chun (2014)^Citation52	No detailed description	Only mentioned randomized	Unclear	0	2
Liu (2014)^Citation53	No detailed description	Only mentioned randomized	Unclear	0	2
Zhang et al (2014)^Citation54	No detailed description	Only mentioned randomized	Unclear	0	2
Shen et al (2014)^Citation55	No detailed description	Only mentioned randomized	Unclear	0	2
Shao et al (2014)^Citation57	No detailed description	Only mentioned randomized	Unclear	0	2
Long et al (2014)^Citation56	Odd or even number	Only mentioned randomized	Unclear	0	1
Tian et al (2014)^Citation66	Randomization chart	Only mentioned randomized	Unclear	0	3
Sun et al (2014)^Citation67	Computer-generated permuted block sequencing	Only mentioned randomized	Unclear	0	3
Yang et al (2013)^Citation58	No detailed description	Only mentioned randomized	Unclear	0	2
Yang et al (2013)^Citation59	No detailed description	Only mentioned randomized	Unclear	0	2
Wu (2013)^Citation60	No detailed description	Only mentioned randomized	Unclear	15.4	3
Huang et al (2013)^Citation24	Computer-generated permuted	Sealed envelopes	Single blinding	0	3
Bai et al (2013)^Citation61	No detailed description	Only mentioned randomized	Unclear	0	2
Zeng et al (2012)^Citation62	No detailed description	Only mentioned randomized	Unclear	0	2
Wang et al (2009)^Citation63	Random-number table	Only mentioned randomized	Unclear	0	3

Notes: The Jadad scale was used to assess the quality of these RCTs. Thirteen studies included had a score of 3 points, which reflected the high quality of the study. The majority of studies had 2 points or lower. These studies can be considered to be of relatively low quality.

Abbreviation: RCT, randomized controlled trials.

Intervention effects

Number of retrieved lymph nodes and metastatic lymph nodes, and metastatic rate of retrieved lymph nodes in the CN and blank control groups

Compared with the blank control groups, the use of CNs resulted in an increased number of retrieved lymph nodes, approximately 3.39 per patient (WMD =3.39, 95% CI =2.73–4.05, P<0.00001; Figure 4A). The number of retrieved metastatic lymph nodes per patient in the CN group was significantly higher than in the blank control group (WMD =0.98, 95% CI =0.61–1.35, P<0.00001; Figure 4B). However, interestingly, the total metastatic rate of the lymph nodes, metastatic rate of the stained lymph nodes, and metastatic rate of the unstained lymph nodes were not significantly different between the CN and blank control groups (OR =1.13, 95% CI =0.87–1.47, P=0.35; OR =1.33, 95% CI =0.91–1.94, P=0.14; and OR =0.55, 95% CI =0.23–1.36, P=0.20, respectively; Figure 5A–C).

Figure 4 Forest plot showing the association of CNs with retrieved lymph nodes and metastatic lymph nodes.

Notes: (A) Number of retrieved lymph nodes per patient in the CN and blank control groups. (B) Number of retrieved metastatic lymph nodes per patient.
Abbreviations: CN, carbon nanoparticle; SD, standard deviation; IV, interval variable; CI, confidence interval; random, random effect; df, degrees of freedom.

Figure 5 Forest plot showing the relationship of CNs and different kinds of retrieved lymph nodes.

Notes: (A) Total metastatic rate of the retrieved lymph nodes. Metastatic rates of the (B) stained and (C) unstained lymph nodes in the CN and blank control groups.
Abbreviations: CN, carbon nanoparticle; M–H, Mantel–Haenszel; random, random effect; CI, confidence interval; df, degrees of freedom.

In the present meta-analysis of these studies, we found that the number of retrieved lymph nodes per patient in the CN group was higher than that of the blank control group. This finding of an increasing number of retrieved lymph nodes corresponds to the improvement in the extent of neck dissection. Further, with increasing removal of dyed lymph nodes, metastatic lymph nodes will also be cleared away simultaneously. Hence, the number of metastatic lymph nodes in the CN group was statistically higher than that in the blank control group. Besides, 4 studies reported that the number of retrieved small lymph nodes (diameter <5 mm) was significantly higher in the CN group.^44,60,61,67 These findings suggest that CNs may help identify tiny, suspicious lymph nodes.

However, it should be noted that the total metastatic rate of the retrieved lymph nodes and the metastatic rate of stained or unstained lymph nodes did not significantly differ between the 2 groups, consistent with the findings of previous studies.^46,68 In Gao and Zhao study,⁴⁶ more metastatic lymph nodes were eliminated in the CN group (6±2.37 vs 4±2.49; P<0.01), but the rate of metastatic lymph nodes did not differ (45.97% vs 47.10%; P>0.05). In the study by Yan et al,⁶⁸ no increase in the number of sentinel lymph node metastasis-positive cases was observed with the utilization of CNs, as compared to that in the control group (36.8% vs 63.2%). Concerning the mechanism of CNs, it is considered that the tissue damage and inflammation caused by the tumor alter the lymphatic drainage channels of the thyroid, which in turn will affect the diffusion of CNs and the identification of metastatic lymph nodes. Thus, it is actually quite hard for CNs to distinguish metastatic lymph nodes among normal lymph nodes.

Anatomic structure and physical function of the parathyroid in the CN and blank control groups

Compared with the blank control group, the use of CNs was associated with a lower rate of accidental parathyroid removal, approximately 22% (OR =0.22, 95% CI =0.16–0.30, P<0.00001; RD =−0.13, 95% CI =−0.15 to −0.11, P<0.00001; Figure 6). The transient and permanent hypoparathyroidism rates declined by approximately 31% and 24% in the CN group, respectively (transient hypoparathyroidism rate: OR =0.31, 95% CI =0.25–0.39, P<0.00001; RD =−0.15, 95% CI =−0.18 to −0.12, P<0.00001; Figure 7A; permanent hypoparathyroidism rate: OR =0.24, 95% CI =0.07–0.85, P=0.03; RD =−0.02, 95% CI =−0.03 to −0.00, P<0.02; Figure 7B). In addition, the rate of postoperative transient hypocalcemia in the blank control group was 30% higher than in the CN group (OR =0.30, 95% CI =0.25–0.38, P<0.00001; RD =−0.16, 95% CI =−0.18 to −0.13, P<0.00001; Figure 8A). On the other hand, the postoperative permanent hypocalcemia rate did not significantly differ between the CN and blank control groups (OR =0.33, 95% CI =0.04–3.03, P=0.33; RD =−0.01, 95% CI =−0.03 to −0.01, P=0.35; Figure 8B).

Figure 6 Accidental parathyroid removal rate in the CN and blank control groups.

Abbreviations: CN, carbon nanoparticle; M–H, Mantel–Haenszel; CI, confidence interval; df, degrees of freedom.

Figure 7 (A) Postoperative transient and (B) permanent hypoparathyroidism rates in the CN and blank control groups.

Abbreviations: CN, carbon nanoparticle; M–H, Mantel–Haenszel; CI, confidence interval; df, degrees of freedom.

Figure 8 (A) Postoperative transient and (B) permanent hypocalcemia rates in the CN and blank control groups.

Abbreviations: CN, carbon nanoparticle; M–H, Mantel–Haenszel; CI, confidence interval; df, degrees of freedom.

In terms of the protection of the parathyroid, some previous studies showed that the usage of CNs was not beneficial;^18,24 however, our study demonstrated that the rates of accidental parathyroid removal (Figure 6), postoperative transient or permanent hypoparathyroidism (Figure 7), and transient hypocalcemia (Figure 8A) were lower in the CN group. This finding suggests that the usage of CNs will help distinguish and preserve the parathyroid glands. However, there was no significant difference in the rate of permanent hypocalcemia. This might have resulted from the quantitative restrictions of the RCTs included. In addition, Yang et al reported that the use of CNs resulted in a decreased rate of parathyroid auto-transplantation.⁴⁷

Publication bias and limitations

A funnel plot analysis of all RCTs was performed as part of the present meta-analysis. The findings indicated that the publication bias was low (Figure 9). But considering the difficulties in publishing studies with negative findings, many studies likely remain unpublished, and these were not available for analysis. In addition, the quality of assessment scores of the 47 RCTs included was relatively low, owing largely to a lack of blinding; however, it is difficult to apply double-blinding during surgery. Thus, further high-quality research is needed to verify the conclusions of the present study.

Figure 9 Funnel plots for publication bias.

Notes: Publication bias for (A) the number of retrieved lymph nodes per patient, (B) number of retrieved metastatic lymph nodes per patient, (C) total metastatic rate of the retrieved lymph nodes, (D) metastatic rate of the stained lymph nodes, (E) metastatic rate of the unstained lymph nodes, (F) accidental parathyroid removal rate, (G) postoperative transient hypoparathyroidism rate, (H) postoperative permanent hypoparathyroidism rate, (I) postoperative transient hypocalcemia rate, and (J) postoperative permanent hypocalcemia rate in the CN and blank control groups. Each point represents a separate study for the indicated association.
Abbreviations: CN, carbon nanoparticle; SE, standard error; MD, mean difference; OR, odds ratio.

Conclusion

This systematic review and meta-analysis demonstrated that the usage of CNs can improve the extent of neck dissection and preserve the normal anatomic structure and physiological function of the parathyroid. At the same time, the number of retrieved metastatic lymph nodes can also be improved during thyroid cancer surgery.

Disclosure

The authors report no conflicts of interest in this work.