The Current Surgical Strategies of Extra-pelvic IVL
Literatures focused mainly on developing an adequate surgical strategy [6-7, 10], but often negate the cost efficiency. Two-stage surgical resection was proven to be effective, but it increased patient burden physiologically and financially. It was true that sterno-laparotomy provided adequate access to the whole path of tumor, however, patients would not be satisfied with such a long incision from suprasternalis to symphysis pubis. Although deep hypothermia circulatory arrest provided a perfectly bloodless field, patients were at risk of related complications such as organ dysfunction, severe coagulation disorder and neurological damage. There were some individual less invasive attempts [15-20] and a few reports[6-7, 10] have proposed surgical guidelines covering part of these strategies, however, in our opinion, these proposals were not ideal and excluded some patients who would have received less invasive surgery. The present study introduces 4 types of surgical strategies (Fig. 2), and try to propose a more comprehensive guideline (Fig. 1) which can accommodate more patients receiving the less invasive surgery without compromising the curative effect.
Single Laparotomy: Tumor Thrombectomy without CPB (Type 1 Surgery)
A few surgical teams had successfully extracted the freely movable tumor from short venotomy without CPB [10, 15, 17-18]. If maximal diameter of the proximal tumor is less than or equal to the diameter of vein at the proximal side of short venotomy (Fig. 2A), it is technically feasible to blindly pull out the proximal length of tumor from short venotomy (Fig. 2A and 3C) through the same midline laparotomy as planned panhysterectomy and bilateral adnexectomy. Bleeding can be confidently controlled by vessel loops around the venotomy, therefore CPB can be avoided. Chest-opening is not necessary for ICL. This criterion excludes the possibility of remote tumor adherence. Blunt dissection of adherence at the CIV or the inlet of ovarian vein is confident through these short venotomies. Caution should be taken in pursuing a long incision length of venotomy, due to the risks of excessive bleeding from drainage of lumbar veins and hemodynamic instability by occlusion of the IVC. For such scenario, CPB should be established.
Single Laparotomy: Tumor Thrombectomy with CPB (Type 2 Surgery)
If the maximal diameter of the proximal tumor is larger than the diameter of vein at the proximal side of short venotomy (Fig. 2B), it is technically impossible to drag the tumor out of short venotomy, an extensive venotomy to the site of dilation is needed to release the tumor (Fig. 2B and 4D). In this category, intracaval tumor circular adherence can be problematic. According to a review of 194 patients , in 76 patients with adherence situation available, 55 patients developed tumor adherence with intima of IVC or heart, it was in accordance with our cohort (50%). Our observations showed that adherences to the wall of extrapelvic veins only (but not always) occurred at the site of the oversize tumor or near the starting site of tumor afflux (Table S1). The mechanism might be the close contact between the surfaces of tumor and intima under compression stress. Pre-operative assessment should pay attention to possible tumor adherences in the middle of IVC. Recklessly dragging a remote tumor adherence belong the short venotomy is at risk of tearing the IVC, an extensive venotomy to the site of adherence is necessary for blunt dissection. The facilitation of CPB is highly recommended in order to reduce risks of hemodynamic instability and blood loss . In this category, if the maximal diameter of intra-cardiac head of tumor is less than or equal to the diameter of the inlet of IVC (Fig. 2B), the intra-cardiac tumor can still be blindly pull downward through the inlet of IVC without chest opening.
The upper end of extensive venotomy could be set at the site of the maximal tumor dilation (Fig. 2B), ranging from the infra renal IVC to the supra hepatic IVC. The retro or supra hepatic IVC could be accessed by mobilizing the right hepatic lobe. After pulling out the proximal head of tumor, a serial and sequential cavotomy could be done until the extraction of tumor out of IIV or ovarian vein. By doing so, the whole length of IVC could be safely opened, any kind of intracaval tumor dilation or adherence at any site of IVC, part of which once was thought to require sternotomy[6-7, 10], can be resected through laparotomy only.
In order to provide well visualization, deep hypothermia circulatory arrest were often applied by other surgical teams [6-7, 10, 14]. In our experience, satisfied bloodless field could be obtained by umbilical tapes encircling around the venotomy together with 2 intracardiac suction devices of CPB , also by the minimal caval opening during serial and sequential cavotomy. Mild hypothermia was approved to be competent for various types of surgeries in our cohort. Therefore, the adverse effects of circulatory arrest and deep hypothermia could be avoided.
After the implementation of the new guideline, 9 of 14 consecutive ICL cases accomplished single-laparotomy strategy in our cohort, with or without CPB. Also, by review of the first 6 ICLs who received sterno-laparotomies indiscriminately, 4 of them fitted these criteria and could have applied single laparotomy strategy. It is encouraged that our guideline accommodated more patients avoiding chest opening compared with previous proposals [6-7, 10].
Double-incisions: Mini-thoracotomy (Type 3 Surgery) or Sternotomy (Type 4 Surgery)
Once the maximal diameter of intra-cardiac head of tumor is larger than the diameter of the inlet of IVC (Fig. 2C), atriotomy is inevitable. In our experiences, if the tumor head congested the intracardiac cavity and showed signs of immobilization by echocardiography (Fig. 6A) massive tumor adherence could be strongly predicted, sternotomy was neccessary to gain adequate exposure. If not, tumor can be confidently resected through right mini-thoracotomy (Fig. 5D,E). Possible lesion of tricuspid valve  could also be repaired or replaced through this minimal incision. Brutal pulling of tumor through atriotomy may tear the intraperitoneal vein , the upward removal of tumor can be safer by firstly transecting the tumor through intra-abdominal venotomy (Fig. 2C).
In our cohort, both cases extending into the PA developed adherences within PA. According to the literatures, adherences inside PA might be present or not present. Blind extraction from venotomy inside abdomen seems unsafe for this category of patients, sterno-laparotomy should be prudently planned considering the need of pulmonary arteriotomy.
There are other important rules. Any blind pulling of intracardiac tumor should be under surveillance of transesophageal echocardiography. The more radical strategy should always be regarded as a backup plan for any lesser invasive strategy. For example, CPB should be on standby during non-CPB thrombectomy. Single laparotomy should be converted to double-incision if blind extraction of intracardiac portion fails.
The Extent of Surgery within Pelvic Cavity
IVLs sometimes diffusely enter the pelvic venous plexus, radical thrombectomies before any ligations of the ovarian and parametrial venous plexuses are necessary, a meticulous search for tumor thrombus is also needed. Even so, some slim tumor thrombus may hide deeply inside the pelvic venous plexus, which raise uncertainty about whether a complete resection is accomplished. Therefore, prophylactic bilateral ligations of IIV and ovarian vein are imperative to prevent potential recurrence entering the IVC [24-25].
In some reports [8, 26-27], hormonal therapies were effective to prevent recurrence after incomplete resections. In our cohort, tumor did not recur for almost 4 years each in the 2 cases following incomplete resections of pelvic intravenous tumor, while hormonal therapy was absent or given for a short period of one year. The role of hormonal therapy is still controversial [5, 9].
This study is a retrospective analysis, the small number of patients analyzed and various confounding factors limits the statistical power. Because of the extreme low incidence of this disease, in our cohort, there was only individual or no case fit for some criteria of the proposed guide which was designed to include all possible forms of extra-pelvic IVLs. More cases are needed in future study to validate the utility of the proposed guide.