Modified sonoelastographic scale score for lymph node assessment in lymphoma – a preliminary report

Abstract

Elastography is a new method of assessment of lymph node consistency. The majority of papers focus on metastases detection of head and neck or breast tumors. The typical desmoplastic reaction in connective tissue stroma in cancer, which is responsible for tissue's hardening, is seen in lymphoma less frequently.

Material and methods

Study of 15 patients with active Hodgkin and non-Hodgkin lymphomas and 16 previously treated patients with no evidence of recurrence. A total of 60 patients suffering from reactive lymphadenopathy was the control group. The size, appearance, vascularity and elastogram of lymph node was analyzed.

Results

Type C elastogram correlated strongly with lymph nodes which presented at least three pathologic features in 2D image. Large lymph nodes with long axis diameter over 3 cm may have ‘cheese with holes’ appearance rather than common type C elastogram.

Conclusion

According to our simplified sonoelastography scoring system, type C elastogram was present mainly in patients with active lymphoma affecting lymph nodes, which showed all four features that we analyzed using conventional ultrasonography. This elastogram was also present in reactive lymphadenopathies. Distinctive for large lymph nodes in patients with active forms of lymphoma “cheese with holes” appearance, can be considered as equal with type C image, although it is corresponding to type B sonoelastographic map.

Introduction

The significance of lymph node consistency on palpation was introduced by Hippocrates in Peri adenôn (Gr. On glands)⁽¹⁾. Neoplastically changed tissue has greater consistency than in health. It is caused by desmoplastic reaction, which is a massive, neoplastic cytokine-induced myofibroblast proliferation with increased collagen and other components of extracellular matrix production^(2–4).

For centuries, palpation was the only method of assessment of desmoplastic reaction progression. It has changed at the turn on XIX-XX century, when microscope technology enabled more precise cell evaluation⁽⁵⁾. Physicians used these two methods solely by the end of the last century.

In 1991, for the first time, Ophir et al. used the term ‘elastography’⁽⁶⁾. The essence of this method is Young's modulus, which measures the ratio of deformation upon the stress being put⁽⁷⁾.

Elastography of lymph nodes was initially based on quasistatistic analysis⁽⁸⁾. The biggest drawback back then, was an inability to evaluate elastographic and 2D images simultaneously. Lyshchik, relying on that, reported high specificity (98%) and sensitivity (85%) for examining lymph nodes in detection of metastases of head and neck tumors, for cutoff point >1,5 for the lymph node-to-muscle strain ratio (SR)⁽⁸⁾. A few years later Tan confirmed these results⁽⁹⁾.

Additionally, Lyshchik retrospectively observed that accuracy of the examination is altered by ultrasonographer's experience, ranging from 79–92%. Over time, more advanced types of elastography appeared, for example real-time elastography (RTE) and shear wave elastography (SWE). RTE depicts consistency of tissue in ROI (region of interest) on color map, where usually ‘warm’ (red, yellow) colors correspond to soft, while ‘cool’ (blue) to hard areas. Having installed additional software, RTE visualize relative strain distribution on condition that a reference tissue (for example muscle, adipose tissue) is within the ROI⁽¹⁰⁾. On comparison, SWE measures exact absolute value of consistency (not a relative strain distribution as in RTE) in kilopascals (kPa)⁽¹¹⁾. SWE probe produces an acoustic radiation force (‘virtual finger’) that deforms tissue and creates shear waves, which velocity is directly dependent on tissue consistency and is based on Young's modulus⁽¹²⁾. In 2014, Youk proved SWE to be of more diagnostic value than RTE (BI-RADS IV in the breast)⁽¹³⁾.

RTE sonoelastography in addition to computing the lymph node-to-surrounding tissue elasticity ratio can create a color map of elasticity. In order to make this new modality useful in reading data, several authors established sonoelastographic scale scores for lymph node assessment.

Most common and currently used scales are the following: Alam⁽¹⁴⁾, Tsukuba (Ishibashi)⁽¹⁵⁾ and Furukawa score⁽¹⁶⁾ (Fig. 1). Recently, Dudei et al. fused wisely sonoela stographic features with traditional 2D image⁽¹⁷⁾. Assured advantage of sonoelastographic color map over RTE-based estimation of consistency ratio is shortening of the duration of examination. Above mentioned statement is applied on our own experience in diagnosing lymphoproliferative disorders, where a high number of enlarged lymph nodes was reported. However, Deng-Ke Teng⁽¹⁸⁾ and Zhi et al.⁽¹⁹⁾ proved deformation ratio to be more credible than scale scores using elasticity colormapping only, yet their studies were based on lymph node metastasis detection. Last but not least, the fact in favour of color mapping is that many companies do not facilitate (or did not so) ultrasound devices with software enabling measurement of deformation ratio. As in our case, when ultrasound machine Aplio 500 bought at the turn of 2011/2012 was featured only with a color map elastography, while SR option was acquired several months later.

Fig. 1.

Comparison of sonoelastographic scale scores for lymph node and image of lymph node according to Alam score and Tsukuba score. Additional images for Tsukuba score

There are dozens of different sonoelastographic scores based on color map consistency of lymph node^(14–16). These scores were created to serve in lymph node metastasis from squamous cell carcinoma of the head and neck region or breast detection. The most common scales in use are Tsukuba score (referred to as Ishibashi score), Alam score and Furukawa score^(14–17).

As far as our research is concerned, there is no paper available on sonoelastrographic scale score used for lymph node assessment in lymphoma.

In contrast to carcinoma, in the majority of non-Hodgkin lymphomas (NHL), desmoplastic reaction is absent or limited if present. In these disorders, consistency of lymph nodes is determined by dense ‘folding’ of rapidly proliferating cells, especially in aggressive type of NHL, for example in Burkitt's lymphoma^(20–23). Exceptions are the advanced subtypes of NS (nodular sclerosis) and MC (mixed-cell) of Hodgkin's lymphoma, where desmoplastic reaction is usually observed.⁽²⁴⁾

The TNM system is commonly used in cancer staging. If any metastasis in lymph node is present, a disease is in stage III or IV (for example papillary thyroid carcinoma). Som demonstrated, that in case of the head and neck region carcinoma, presence of metastasis in lymph node decreases 5-year survival rate for 50%, and in case of contralateral metastasis for 75%⁽²⁵⁾.

The number of neoplastically altered lymph node sinlymphoma is not so important as in the TNM system. In lymphoma, a focus is put on the topography of changes to the diaphragm and as for staging, AnnArbour Staging for Hodgkin's disease and NHL is used in adults⁽²⁶⁾ and Murphy Staging for NHL in children^{(27, 28)}. Bearing in mind pathomorphologic discrepancies and different staging system, we decided to conduct sonoelastographic study and create a scale score for lymph node assessment trying to make the existing ones more accessible. Examination of superficial lymph nodes were carried out as follows: the neck (Rouvière's regions I–IV)⁽²⁹⁾, the axilla (I region), the infraclavicular fossa, the groin and parasternal lymph nodes.

Patients and methods

Ultra sonographic examination was performed with UIM V-Aplio500A device, including RTE sonoelastography, using a PLT- 805AT probe with a frequency of 12 MHz.

The research group aged 10 – 67 was made up of: 15 patients presenting active disease (HL – 9 i NHL – 6), 16 previously treated patients with no evidence of recurrence (HL – 6 i NHL – 10). A total of 60 patients suffering from reactive lymphadenopathy was the control group. 2 patients suffering from HL were excluded as chemotherapeutic treatment was applied before the examination. Sonoelastography of 243 lymph nodes and lymph node packages in patients presenting active form of lymphoma, 16 lymph nodes in patient presenting lymphoma in remission and 202 reactive lymph nodes was carried out.

The procedure of the study consisted of ultrasonography of the superficially mph nodes with linear probe (12 MHz) – located in the neck (regions I–VI), axillae (region I), infraclavicular fossae, groins and parasternally – and the reading of images regarding: its standard size according to localization ⁽³⁰⁾ (Tab. 1), III^rdor IV^thWesthofen's vascularisation type⁽³¹⁾ (Fig. 2), cortical echogenicity decrease and the lack of hyperechogenic lymph node sinus. Sonoelastographic a ssessment of lymph nodes fulfilling two of a forement ioned criteria and directly adjacent lymph nodes (even with normal ultrasound image) was made. Lymph node packages showing polycyclic shape were regarded as ‘a single lymph node’. We did not differentiate neoplastically changed lymph nodes from reactive ones, though it was feasible in children presenting HL – all patients were included in experimental treatment protocol (EuroNet-PHL-C1), according to which a PET- CT scan was routinely performed before the treatment (approximately together with sonoelastography). We decided not to include results of PET-CT scans of patients with NHL and adults, which were not performed routinely according to ESMO guidelines (European Society for Medical Oncology)⁽³²⁾.

Tab. 1.

Standard size of lymph node according to localization

Localization	Ultrasonographic standard size
Neck*	<5–8 mm – short axis(37)
Supraclavicular fossa	<5 mm – short axis(38)
At the angle of mandible** (jugulodigastric and omohyoid nodes)	<15–20 mm – long axis(39, 40) ≤9 mm – short axis(41)
Parenchymal parotid lymph nodes	<5 mm – short axis(37)
Axilla and infraclavicular fossa (levels I–III)	<20 mm – long axis and <10 mm – short axis (42)
Parasternal	<6 mm – long axis(43)
Groin	<15 mm – short axis(42) <20 mm – long axis(44, 45)

^*Due to high prevalence of cervical lymphadenopathy in children (90% in children aged 4–8 according to Park) some sources (Cummings CW: Otolaryngology: Head and Neck Surgery) imply that an enlarged lymph node in children means lymph node >2 cm in long axis.

^**These lymph nodes are classified as II^nd cervical nodular level, yet they are permanently enlarged reactively as they drain the majority of lymph from tonsillar region and oral mucosa

Fig. 2.

Westhofen's lymph nodes vascularisation types

The examination consisted of multiple pulsating pressure, that was perpendicular to the skin over nodular areas, so as compression curve was sinusoid-shaped with ‘rounded peaks’ (Fig. 3).

Fig. 3.

A series of images of the same lymph node in the popliteal region in patient who has undergone enucleation of knee joint and prosthesis implantation due to sarcoma. Series of images A: Inappropriately performed examination presents different image at point of maximum compression than at baseline (see also: Fig. 4). Note a spiked wave of tissue compression curve. Series of images B: Examination of the same region and patient performed properly presents similar images both at the peak of the curve and at baseline. Parabolic shape of tissue compression curve

Elastograms in properly performed examination were similar in crossing points of the baseline and points of ma ximal compression and decompression (Fig. 4A). Otherwise the study was repeated, as it indicated insufficient decompression of tissue. When thickness of soft tissue was smaller (for example assessing nuchal lymph nodes), we were satisfied of plane sinusoid-shaped curve. In those cases we interpreted images in their point of maximal compression and transition through the curve on time of compression, as color contrast of lymph nodes in these type of images was better than in those obtained on time of decompression (Fig. 4B). Several authors⁽³³⁾ claimed that evaluation at the point of maximal compression is also of diagnostic value, yet nowadays assessment during decompression is recommended.

Fig. 4A.

Types of compression-decompression curves. 1 – Symmetric sinusoid with similar velocity of compression and decompression increase and similar height of waves during compression and decompression. As for this type of curve, almost the same images at point of maximal compression, maximal decompression and at crossing point of baseline can be obtained. 2 – Spiked wave asymmetry, velocity of compression increase is greater than of decompression, similar height of waves during compression and decompression – similar result were obtained only at crossing point of baseline during maximal compression. 3 – Spiked wave asymmetry, velocity of compression increase is smaller than of decompression, similar height of waves during compression and decompression- similar result were obtained only at crossing point of baseline during maximal compression. 4 – Flat-shaped curve asymmetry with sharp waves of maximal compression and decompression. Reproducible results only at point of maximal decompression. 5 – Flat-shaped curve asymmetry with flat waves – results not reproducible. 6 – Curve asymmetry with present compression-decompression arrhythmia – results not reproducible

Fig. 4B.

Superficial lymph nodes ultrasound image, when sinusoid-shaped curve of compression cannot be obtained. Image A – sonoelastogram at point of maximal compression of plane curve; image B – sonoelastogram at point of maximal decompression. Different levels of elasticity in lymph node is more visible during compression, while differentiation of elasticity between lymph node and surrounding tissue is more visible on decompression

The evaluation was based on created and used in our center modified sonoelastographic scale score for lymphoproliferative disorders, which is a chimera of Tsukuba and Furukawa scores (Fig. 5).

Fig. 5.

Modified sonoelastographic scale score for lymph node assessment in malignant lymphoproliferative disorders

Results

One previously treated patient was diagnosed with a extensive residual mass in the mediastinum (medical history: PET-CT scan result), which was partly seen on 2D ultrasonography scans. RTE was not reproducible, thus we excluded that patient from our study.

All four features of lymph node in 2D ultrasound considering patient s presenting active form of a disease were shown in 66 cases, 3 features in 119 and 2 features in 58 cases. In ours modified sonoelastographic scale score (Tab. 2) feature A corresponded to lymph nodes presenting >50% of warm color (green, red), B – >50% of cool color (blue), feature C corresponded to the V^th grade in Tsu kuba's score, where surrounding tissue was infiltrated or the image was of “cheese with holes” appearance (Fig. 6). The “cheese with holes” appearance is our own observation that was not mentioned before. This image corresponded to Tsukuba/Furukawa IV^th grade and was reproducible in crossing point of baseline during decompression and was characterized by small areas of ‘soft colors’ in crossing point of baseline during compression. These areas represented blood vessels what was confirmed in power Doppler examination (Fig. 6). The results are as follows: 4 features: type A elasto gram – 10; type B – 47, type C – 9; 3 features: type A – 18, type B – 96, type C – 5; 2 features: type A – 16, type B – 40, type C – 2. Type C elastogram correlated strongly with 4 and 3 features present, 0,72 and 0,51 respectively. For 2 present features the correlation was weak (0,16). Type C elastogram was also present considering reactive lymphadenopathy, but with low frequency for 2 features – 0,7%; for 3 features the frequency was surprisingly high – 10,3%, what was probably the result of low number of cases (<30 elements). It was not observed in previously treated and in remission patients. For type C we assumed lymph nodes that presented grades III–IV only, according to Alam's or Tsukuba's sonoelastographic scale scores.

Tab. 2.

The incidence of lymph nodes presenting 2, 3 and 4 pathologic features in 2D ultrasound and elastograms according to modified scale score for lymphoma

2 features
	Active HL/NHL	HL/NHL in remmision	Reactive lymphadenopathy
A	16	9	86
B	40	7	53
C	2	0	1
Total	58	16	140
3 features
A	18	0	18
B	96	2	9
C	5	0	3
Total	119	2	29
4 features
A	10	0	14
B	47	0	4
C	9	0	0
Total	66	0	18

Fig. 6.

Extensive nodular mass (5,3 × 2,4 cm) in right supraclavicular fossa in patient presenting HL. Pronounced type III vascularisation, lack of fatty hilum and areas of heterogeneously decreased echogenicity. At crossing point of baseline on decompression, a lymph node presents type IV, whereas on compression-type III/IV of Tsukuba image. “Cheese with holes”-like areas correspond to visible flow on power Doppler

Discussion

Lymphoma requires distinct diagnostic algorithmas it presents specific for malignant lymphoproliferative disorder features. Although the methods of choice in staging are PET-CT and MR modalities⁽³⁴⁾, ultrasonography shows the highest level of sensitivity reaching 96,8% (for PET-CT – 91,8%). However, even with the use of FNAB (fine needle aspiration biopsy), specificity of ultrasound examination is not as high as in other modalities. Therefore finding new methods, such as sonoelastography, can attribute to specificity level improvement. Desmoplastic reaction is limited and rarely observed in NHL's, LP-HL's (nodular lymphocyte predominant Hodgkin's lymphoma) and LD-HL's (lymphocyte depleted Hodgkin's lymphoma). The rapidly proliferating cells’ ‘pressure’ is in charge of lymph node's consistency. Desmoplastic reaction is seen in NS (nodular sclerosis) type and MC (mixed-cell) type of Hodgkin's lymphoma, predominantly in advanced stages of the disease, yet in cellular phase of nodular sclerosis may be absent^{(35, 36)}. This feature of NS-HL differentiate them from carcinoma metastasis in lymph node. Considering the above, together with the fact of repeated ultrasound examinations of superficial nodular regions, simple and specific sonoelastographic scale score for lymphoma emerges.

According to the authors, an appropriate tissue compression technique is required to use ours and others sonoelastographic scale scores properly. This is the only way to obtain reproducible elastograms.

Although the study is still in progress, the preliminary results are encouraging for carrying them on. We would like to ask other clinical centers for cooperation and verification of practical effectiveness and usefulness of the above proposed scale score.