“The use of ultrasound in cancer diagnosis is not new. However, researchers have, for the first time, been able to show that patterns of ultrasound signals can be used to identify whether or not cancer has started to spread in melanoma patients, and to what extent. They believe the discovery will help doctors determine the amount of surgery needed and better reach a prognosis for the patient.”
The above is taken from the European Cancer Organisation meeting, held jointly with the European Society for Medical Oncology, which took place in Berlin in 2009. But what does it mean in detail?
Melanoma is by far the most aggressive skin cancer known. Its incidence and to a lesser extent the mortality rate has increased over the last 60 years by about 700%. The reason why the mortality rate has only slightly increased is based on the predominant increase of thin melanomas with a good prognosis, which do not yet show distant dissemination. Awareness campaigns might have contributed to this more frequent early diagnosis of skin cancer. Melanomas occur de novo, or in pre-existing nevi mostly in sun-damaged skin with typical patterns of mutations in oncogenes and even in acral and mucosal areas, here mostly connected with c-kit mutations.
Early clinical observation studies in cancer patients suggested that tumours spread in a methodical, stepwise fashion from the primary site, to the regional lymph nodes, and only then to distant locations. Based on these observations, the regional lymph nodes were believed to be mechanical barriers, at least temporarily preventing the widespread dissemination of tumour (barrier theory, linear spread). Therefore if metastases are detected early enough, it would lead to a better prognosis and with that to a better overall survival rate. Following this theory it has also been observed that most melanomas follow an orderly progression and predominantly spread via the lymphatic ways to the lymph nodes. From the lymph node the disease may progress to distant organs by hematogenous dissemination or may stay restricted to lymph nodes.
Alternatively or in addition, a second way of dissemination is assumed. Distant metastases are said to develop from hematogeneously disseminated tumour cells that originate from the primary tumour and cannot be stopped if genetically planned (parallel spread; “seed and soil” – theory). Thus it is concluded that the number of positive lymph nodes is therefore simply a marker, but not a cause of distant metastases and treatment of lymph nodes would not affect long-term survival. This gene-based control of metastases implies a principal process of metastatic spread for solid tumours. The hypothesis “metastases do not metastasise” has a high plausibility. Following this hypothesis, reduction of lymph node dissection and above all its performance only in those cases where it is necessary for treatment decisions seems to be (bio)-logically consequent.
Following the first theory an elective lymph node dissection (ELND) was performed prophylactically in former times in all melanoma patients with a certain tumour thickness. However, this did not contribute to a general survival benefit or only for subgroups, as could be proven.
Thus it was quite a great step from an elective lymph node dissection to a lower grade surgical invention, which was introduced by the use of sentinel node biopsy (SLNB), which is now used worldwide following the primary diagnosis of a melanoma above 1mm tumour thickness.
Sentinel lymph node biopsy (SLNB) using the triple technique with preoperative lymphoscintigraphy, gamma counter and intra-operative blue-dye has become standard care in the staging of melanoma patients with a primary tumour above 1mm thickness in many countries. So far it is a diagnostic tool to predict survival, yet there might be a survival benefit for a subgroup of SLNB-positive patients.
However, 80% of all patients are somewhat over-treated even by a SLNB because only 20% of all SNs will be metastatic. Moreover, most patients have clinically undetectable micrometastases, some even so-called submicroscopic tumour burden (metastases within SN < 0.1mm in diameter). The SN sample and minimal SN metastatic disease when defined as isolated clusters of cells warrant further study as a potential indicator against CLND after positive SLN. Moreover, it could be shown that a different lymph node involvement based on the SN tumour burden reflects a different type of disease. Thus gross lymph node involvement correlates almost invariably with very poor prognosis and a small subcapsular metastasis can be prognostically as favourable as a negative sentinel node diagnosis. Thus, in summary, in melanoma patients with positive SLNs, SN tumour burden, primary tumour thickness, and the number of SNs harvested may be useful in identifying a group at low risk for positive non-SLNs that can be spared the potential morbidity of a completion lymph node dissection (CLND).
For several years our Berlin-Rotterdam Research Group has been interested in the use of ultrasound during the follow-up of melanoma patients, the use of fine needle aspiration cytology (FNAC) for verification of suspected lesions and the additional application of molecular-biological methods out of in vivo, i.e. FNAC material. Its goal was to test if the SN can be reliably depicted by ultrasound in vivo, meaning that the SLNB was still in the patient prior to SLNB operation, and if a tumour involvement can subsequently be verified by FNAC.
Since we could prove that ultrasound can reliably detect the excised SN prior to excision with a high sensitivity in the pilot study, our technique could also examine other potential lymph nodes at other sites at risk for metastases. The first prospective data regarding 400 consecutive patients demonstrated that ultrasound-guided FNAC could identify 65% of all SN metastases prior to the surgical excision of the SLNB. The rate of the missed patients by ultrasound-guided FNAC in the total population (including SN negative patients) was 28/400 (7%) and herewith corresponds to about the same false negative rate of the SLNB procedure. Moreover, the survival curves for disease-free as well as overall survival for the patients staged with ultrasound and FNAC and for the patients staged with regular SLNB mimic each other and are therefore in favour of our non-invasive method.
The advantage of the present approach is that it is more minimally invasive than the SLNB and saves the patients with proven involvement in FNAC an SNLB and thus a surgical intervention as patients can proceed directly to the CLND. For the future it could indicate that only the patients with a suspicious ultrasound and a lack of verification by FNAC would undergo a SLNB.
In case the ultrasound-guided FNAC method overlooked an involved SN, this can easily be detected during follow-up in regular intervals as could be proven in other possible future studies. It was shown in this study that the SUB-micrometastases will regularly be overlooked by ultrasound. However, in a further own multi-centre study it was also demonstrated that SUB-micrometastases < 0.1mm have the same prognosis as SN negative patients. So those “overlooked” patients were likely not to be harmed by performing ultrasound-guided FNAC, although all patients in the study did proceed to undergo a CLND.
The disadvantage of ultrasound-guided FNAC is the relative dependency on the examiner. This means that all sonographers will have a learning curve. To make this method more accessible to every examiner we performed another study. The background of this study was the experience gained throughout the years in the detection of morphological patterns. Moreover, an important factor to perform this study was the lack of clear-cut morphological patterns in the ultrasound examination of melanoma lymph nodes, unlike breast or thyroid cancer, which have clearly defined, but different patterns. Therefore this study examined a pre-determined set of morphological criteria to the golden standard of histopathology.
All morphology results were correlated to the SLNB tumour burden measured according to the Rotterdam Criteria and the location of the tumour nests within the SN (subcapsular vs intraparenchymal) thus providing us with a correlation of the pattern and the grade of involvement.
Thus we could even observe a development “over time” of certain patterns from the way of an early involvement to a late metastasis. Mainly, peripheral perfusion (PP) is observed as an early sign of metastasis. However, this sign remains in later phases of metastases as irregular perfusion of the entire node. This reflects a process by which cancer cells in the lymph node are starting to set up a support system, involving new blood vessels to supply them from outside the lymph node. This tends to happen on the edge of the lymph node where the cancer cells are clustering to form the hump that the ultrasound sees.
The loss of central echoes (LCE) reflects a more dynamic process. At first a displacement of central echoes to the periphery can be observed (this is a later sign, as already a significantly large metastasis has developed). Later on, when the metastasis has grown even larger, the central echo sign is completely lost (total involvement by tumour). This finally ends up in a balloon-shaped lymph node (BS) similar to the total loss of the central echoes meaning that a tumour has mass involvement of the entire lymph node; thus relatively most / all of the lymph node has become involved by a metastasis and thus the lymph node in abnormally enlarged and loses its normal long and slim appearance.
SLNB provides the best information on prognosis and provides us with an important tool to stratify for and study more homogeneous patient populations to evaluate adjuvant systemic therapies in randomised phase III trials. The utility of systemic adjuvant therapy remains marginal as a result of the fact that a lack of effective drugs in stage IV disease is reflected by a lack of effective adjuvant therapies in stage II-III melanoma. Thus far, chemotherapeutic drugs, immunostimulants, and various vaccines have all failed. Interferon (IFN) has an effect on relapse-free survival, but not on overall survival.
New data suggest strongly that RFS as well as OS can be favourably impacted by adjuvant IFN or pegylated IFN therapy in patients with an ulcerated primary and minimal tumour load (sentinel node positive) in the regional lymph nodes. This means that identifying a positive sentinel node in melanoma patients may have consequences for both surgical therapy and systemic therapy decisions
SLNB, however, also means an overtreatment for about 80% of the melanoma patients undergoing an SLNB that will have a negative histopathology. On the other hand SLNB is a surgical procedure, minimally invasive, but still a procedure with complications, morbidity and costs. Ultrasound-FNAC is less invasive, is virtually without complications and morbidity and might well be a cost-effective approach.
In short, this research could change clinical practice enormously. If the applicability of these US patterns can be reproduced in a prospective multi-centre trial as is foreseen within the framework of EORTC Melanoma Group, the unselective SLNB approach in all patients with melanomas thicker than 1mm could be replaced by a more selective approach by offering it to the patients at higher risk for relapse, in other words for those with a suspicious ultrasound pattern.